Week 1 HW: Principles and Practices

1. The Application: AI-Powered Science Automation

I’m interested in building an AI platform that helps automate parts of the scientific process — things like scanning literature for gaps, designing experiments, running them through lab robots (like the Opentrons we’ll use in HTGAA), and helping write up results.

Why? Science is slow. Not because scientists are lazy, but because there’s way more good questions than people to work on them. Many ideas never get tested because the person who had them didn’t have the right lab skills or equipment. And honestly, a lot of published research can’t even be reproduced because of human error in complicated protocols. Or negative results don’t get published at all, leading to the “chasing the same dead ends” phenomenon — but no one knows, because it’s not published.

An AI platform could help with all of that. Not by replacing scientists, but by letting more people do better science faster, use negative and positive results to iterate faster and learn from more data, which can be used to train the next “physics” model of the AI. I think of it like a student somewhere without access to a fancy lab — they could design a CRISPR experiment, have a robot run it remotely, and get solid results back. OpenAI did something very similar now with Ginkgo Bioworks, read here: GPT-5 Lowers Protein Synthesis Cost.

4. Scoring Matrix

Scale: 1 = best, 3 = least effective

6. Ethical Reflections

Going into this week I thought governance is something you deal with after a technology exists. The recitation changed that — the Jurassic Park meme sounds silly but captures it well. We’re too much in “can we?” mode and not enough in “should we?” mode.

The openness question kept bugging me. My gut says make everything open, but then I think about what “everyone” includes and it gets uncomfortable. I now think openness with checkpoints makes more sense — open tools, but controls where designs become physical (synthesis, robot instructions).

AI-generated fraud was new to me. An AI could make up data that looks real, or accidentally lead someone to design something harmful. Provenance tracking for AI outputs seems necessary.

These discussions are also very US-centric. As a med student in Vienna — AI doesn’t stop at borders. Building safety into the platform architecture could raise the floor globally, similar to how iGEM runs safety reviews across all countries without needing international treaties.

Actions I’d propose: ethics review before new AI capabilities get released, provenance tracking as default, tying capability releases to safety milestones, and building risk education directly into the workflow so users can’t blindly automate dangerous stuff.

Week 2 Lecture Prep

AI Disclosure

Claude (Anthropic) — Used to help structure and refine this assignment. The core ideas and positions are my own.

• Prompt 1: “Help me structure my governance analysis for AI-powered science automation, with three governance actions and a scoring matrix.”
• Prompt 2: “Nice I have done the homework draft now, please refine it so it has less spelling errors, correct my grammar and format it better. If you correct my wording, don’t write AI but write human like. Keep all the info unless it is obviously wrong.”

Cursor (AI-assisted IDE) — Used to build and deploy my HTGAA website.

• Prompt 1: “Help me make a cool website. Name Constantin Convalexius. Contact info: Email: convalexius@gmail.com, LinkedIn: https://www.linkedin.com/in/constantin-convalexius/, X: https://x.com/convalexius. About Me: I’m a medstudent interested in Longevity Biotech, Labautomation. Fill in my profile from other chats in Claude. How can I insert a cool cover pic?”
• Prompt 2: “Upload my homework to the website, format it nicely in Markdown, commit and push.”

Week 2 HW: DNA Read, Write, & Edit

Week 2: DNA Read, Write, & Edit

Student: Constantin Convalexius
Course: HTGAA Spring 2026
Location: Vienna, Austria

Part 1: Benchling & In-silico Gel Art

Butterfly art 1

2nd picture: all enzymes

Part 2: Gel Art - Restriction Digests and Gel Electrophoresis

As a committed listener in Vienna without local wet-lab access, I completed the in-silico design and simulation sections.

Part 3: DNA Design Challenge

3.1 — Protein Choice: PD-L1 (Programmed Death-Ligand 1)

I chose PD-L1 (CD274, UniProt: Q9NZQ7) — the immune checkpoint protein that tumor cells use to hide from the immune system. PD-L1 sits on the surface of cancer cells and binds to PD-1 on T-cells, essentially telling them “don’t attack me.” Drugs like Pembrolizumab (Keytruda) block this interaction by targeting PD-1, so the immune system can recognize and destroy the tumor again. As a med student, this is one of the most exciting developments in oncology I’ve encountered so far.

The full-length PD-L1 protein is 290 amino acids and includes a signal peptide, extracellular domain, transmembrane region, and a short intracellular tail. For this exercise, I’m only using the extracellular domain (AA 19-238, 220 residues), since that’s the part that actually interacts with PD-1 and is the relevant domain for drug binding studies. This is also what researchers typically express recombinantly — you don’t need the transmembrane anchor if you just want to study the binding interface.

Protein sequence (extracellular domain):

FTVTVPKDLYVVEYGSNMTIECKFPVEKQLDLAALIVYWEMEDKNIIQFVHGEEDLKVQH
SSYRQRARLLKDQLSLGNAALQITDVKLQDAGVYRCMISYGGADYKRITVKVNAPYNKINQ
RILVVDPVTSEHELTCQAEGYPKAEVIWTSSDHQVLSGKTTTTNSKREEKLFNVTSTLRIN
TTTNEIFYCTFRRLDPEENHTAELVIPELPLAHPPNERTHLVILG

3.2 — Reverse Translation

I used the Sequence Manipulation Suite (SMS2) reverse translation tool with “most likely codons” to convert the amino acid sequence into a DNA nucleotide sequence.

The output was an 870 bp sequence for the full-length 290 AA protein. One thing I noticed is that the SMS2 tool defaults to E. coli codon preferences — you can see this in the output, which uses codons like CGC for Arginine, GCG for Alanine, and CCG for Proline. These are all heavily biased toward bacterial tRNA pools, which wouldn’t work well in a human expression system.

This step is mainly useful to show the “raw” reverse translation before optimization, and to demonstrate why codon optimization is necessary.

3.3 — Codon Optimization

Since I want to express PD-L1 in human HEK293 cells (see 3.4), I ran the extracellular domain amino acid sequence through GenScript’s GenSmart Codon Optimization Tool with Homo sapiens as the host organism.

Results:

Optimized DNA sequence:

TTCACCGTGACCGTTCCAAAGGATCTGTACGTGGTCGAGTACGGCAGCAACATGACCATC
GAGTGCAAGTTCCCCGTGGAAAAGCAGCTGGACCTGGCCGCTCTGATCGTGTACTGGGAG
ATGGAAGATAAGAACATCATCCAGTTCGTGCACGGCGAGGAAGATCTGAAAGTGCAGCAC
AGCAGCTACAGACAGAGAGCCAGACTGCTGAAGGACCAGCTGTCTCTGGGAAATGCTGCC
CTCCAAATCACCGACGTGAAGCTGCAAGACGCCGGCGTGTACCGGTGCATGATCAGCTAT
GGCGGAGCCGACTACAAGAGGATTACCGTGAAAGTGAACGCCCCTTACAACAAGATCAAC
CAGCGGATCCTGGTCGTGGACCCTGTGACATCCGAGCACGAGCTTACATGTCAGGCCGAG
GGCTACCCTAAGGCCGAAGTGATCTGGACCTCCTCTGATCACCAGGTGCTGAGCGGCAAG
ACCACCACCACCAATAGCAAGCGGGAAGAAAAACTGTTTAACGTGACCAGCACACTGAGA
ATCAATACCACAACAAACGAGATCTTCTACTGCACATTCAGAAGACTGGACCCCGAGGAA
AACCACACCGCCGAGCTGGTGATCCCCGAGCTGCCTCTGGCTCATCCTCCTAACGAGAGA
ACACACCTGGTGATCCTGGGC

The key difference compared to the raw SMS2 output is that GenSmart replaced the E. coli-preferred codons with those matching human tRNA abundance. For example, Arginine now uses AGG/AGA/CGG instead of bacterial CGC, and Alanine uses GCC/GCT instead of GCG. This is important because if the codons don’t match the host’s tRNA pool, the ribosome stalls during translation, leading to low protein yields or truncated products.

The GC content of 55.07% is also nicely within the ideal window — too high or too low GC content can cause issues with mRNA secondary structures or difficulties during DNA synthesis.

The codon-optimized sequence was generated using the GenSmart Codon Optimization Tool [1].

[1] Long Fan (2020, February 6). Codon optimization. (WO Patent WO 2020/024917 A1). Nanjing GenScript Biotech Co., Ltd.

3.4 — Production Technologies

Cell-dependent expression (primary approach): HEK293 cells

PD-L1 is a glycoprotein — it has N-linked glycosylation sites that are important for its folding and function. Because of this, I would express it in HEK293 human cells rather than E. coli. The workflow would be: clone the codon-optimized gene into a mammalian expression vector, transfect HEK293 cells, let them express and secrete the protein (since we’re only using the extracellular domain without the transmembrane anchor, it should be secreted into the culture medium), and then purify it using an affinity tag (like a His-tag with Ni-NTA chromatography). HEK293 cells are well-established for this — they handle human post-translational modifications properly and give reasonable yields.

Cell-free expression (alternative):

For quick small-scale testing (e.g., to check if the construct expresses at all before committing to a full cell culture run), you could use an in vitro transcription/translation system like rabbit reticulocyte lysate or wheat germ extract. These systems can produce protein in a few hours rather than days, but they don’t perform proper glycosylation, so the protein wouldn’t be fully functional. Still useful as a rapid validation step.

Part 4: Prepare a Twist DNA Synthesis Order

Here are my screenshots and files for Homework Part 4:

Upload sequence to Twist

Benchling expression cassette map

Twist clonal gene order configuration

PDF export

PDF version prepared locally (not uploaded in this commit).

PDF update: plasmid map screenshot

Part 5: DNA Read, Write, Edit

5.1 DNA Read

(i) What DNA would I want to sequence?

I’d want to sequence the genomes of supercentenarians — people who’ve made it past 110. These individuals somehow dodge or massively delay the diseases that kill most of us (heart disease, cancer, dementia), and there’s evidence that protective variants in genes like FOXO3, APOE, and TERT are enriched in their genomes. But we probably haven’t found everything yet. By doing whole-genome sequencing on large cohorts and comparing them to people who aged “normally,” we could uncover rare genetic variants that essentially act as nature’s longevity engineering. Pair that with DNA methylation data (which feeds into biological aging clocks like the Horvath clock) and you get a pretty complete picture of both the genetic hand they were dealt and how their gene expression shifted — or didn’t — over time.

(ii) Sequencing technology

I’d go with a hybrid approach: Oxford Nanopore (PromethION) for long reads plus Illumina NovaSeq for high-accuracy short reads.

Nanopore (third-generation): Sequences native, single DNA molecules in real time — no PCR amplification needed, which avoids amplification bias. A motor protein threads a DNA strand through a tiny biological pore in a membrane. Each base passing through disrupts the ionic current in a characteristic way, and a neural network translates those current patterns into sequence. Big advantage: it can also detect DNA methylation directly from the native strand, no bisulfite conversion needed. Reads are long (often >20 kb), which helps resolve structural variants and repetitive regions.

Input prep: Extract high-molecular-weight DNA from blood, ligate sequencing adapters directly — pretty minimal compared to short-read platforms.

Illumina (second-generation): Supplements Nanopore with very accurate short reads (~150 bp) for reliable SNP calling. Input prep involves fragmentation, adapter ligation, and bridge PCR. Bases are called by detecting fluorescent signals from reversible dye-terminators during synthesis-by-sequencing cycles.

Output: Both produce FASTQ files. Together they give you phased, chromosome-level assemblies with both structural resolution and single-nucleotide accuracy.

5.2 DNA Write

(i) What DNA would I want to synthesize?

I’d synthesize an engineered human telomerase (hTERT) expression cassette — a gene therapy construct to transiently reactivate telomerase in adult cells.

Telomere shortening is one of the core hallmarks of aging. Every cell division chips away at the protective chromosome caps until the cell senesces or dies. Telomerase rebuilds them, but it’s silenced in most adult tissues. Maria Blasco’s group at CNIO showed that AAV-delivered telomerase in mice extended lifespan without increasing cancer. The idea is to build a controllable human version.

The construct (~6-7 kb) would include a codon-optimized hTERT coding sequence under a Tet-On inducible promoter (so you can switch it on/off with doxycycline — you really don’t want constitutive telomerase, that’s a cancer risk), plus a GFP reporter to track which cells are expressing it. For Twist, I’d order this as overlapping clonal gene fragments.

(ii) Synthesis technology

Phosphoramidite oligo synthesis (Twist Bioscience’s platform) combined with Gibson Assembly.

Twist synthesizes thousands of short overlapping oligos (~60-200 nt) in parallel on silicon chips. Each oligo goes through cycles of deprotection -> coupling -> capping -> oxidation. These oligos get assembled into longer gene fragments (~1.8 kb) via overlap extension, then cloned into plasmids and sequence-verified. For my full ~7 kb construct, I’d order 3-4 fragments from Twist and stitch them together with Gibson Assembly.

Limitations: Coupling efficiency is _{99-99.5% per step, so errors accumulate with length — that’s why you assemble from short oligos rather than synthesizing one long piece. Extreme GC content or repetitive sequences can cause synthesis failures. Turnaround is 2-3 weeks, and cost is around $0.07-0.09/bp (}$500 for the full construct).

5.3 DNA Edit

(i) What DNA would I want to edit?

Three targets for a “longevity panel”:

1. PCSK9 knockout: People with natural loss-of-function mutations in PCSK9 have very low LDL cholesterol and near-immunity to coronary heart disease — the #1 killer globally. A permanent gene edit would be a one-and-done solution. Verve Therapeutics is already running clinical trials on this.
2. TP53 enhancement: Not a knockout — that would be terrible. Instead, introducing “super-p53” gain-of-function variants (studied in mouse models) that boost cancer surveillance without accelerating cellular senescence. The goal: decouple tumor protection from the aging program.
3. Myostatin (MSTN) partial reduction: Myostatin inhibits muscle growth. Sarcopenia (age-related muscle wasting) is a huge driver of frailty in older adults. Reducing myostatin signaling could help maintain muscle mass well into old age — think Belgian Blue cattle, but a gentler, partial version for humans.

George Church has discussed similar multi-gene longevity editing in the context of GP-write.

(ii) Editing technology

For PCSK9: adenine base editing (ABE) via lipid nanoparticles (LNPs). A Cas9 nickase fused to a deaminase enzyme converts a single A·T base pair to G·C, introducing a premature stop codon in PCSK9 — no double-strand break needed. LNPs are delivered IV and preferentially target the liver (perfect for PCSK9). Verve’s primate data shows >60% editing efficiency.

For TP53 and MSTN: prime editing, which uses a Cas9 nickase fused to a reverse transcriptase guided by a pegRNA containing both the target sequence and the desired edit template. Even more precise than base editing — can make any small substitution without double-strand breaks or donor DNA.

Steps (base editing example): Design a guide RNA positioning the target adenine in the editing window -> formulate ABE mRNA + sgRNA in LNPs -> IV infusion -> LNPs enter hepatocytes via ApoE-mediated uptake -> base editor converts A to inosine (read as G) -> permanent single-nucleotide change.

Limitations: Off-target editing risk (lower than standard Cas9 but not zero — needs WGS validation). LNPs mostly hit the liver, which is great for PCSK9 but not for muscle or systemic edits — those need AAV or next-gen tissue-tropic delivery. Prime editing efficiency is still variable (~5-50%). And of course, these edits are permanent and irreversible, which is both the point and the risk.

AI Disclosure

I used Cursor and Claude to help with formatting, spelling/grammar clean-up, and publishing this website documentation.

Week 3 HW: Lab Automation

Week 3: Lab Automation

Student: Constantin Convalexius
Course: HTGAA Spring 2026
Location: Vienna, Austria

Part 1: Python Script for Opentrons Artwork

I created and tested an Opentrons Python script that generates a dotted skull design for gel art.

1.1 What I completed

• Designed a skull artwork concept and implemented it in Python for Opentrons (apiLevel 2.20).
• Used multi-color patterning with helper functions for safer droplet detachment (dispense_and_detach).
• Simulated the protocol in Colab and fixed simulator compatibility issues (e.g., replacing direct protocol.comment calls with mock-safe logging logic).
• Generated a higher-resolution version of the skull by increasing point density.

Submission status

• Artwork script: completed.
• Opentrons skull design image: completed.
• I will submit the Python script for robot execution as required by the course submission form.

1.2 Proof of Opentrons skull artwork

Part 2: Post-Lab Questions

2.1 Published paper using Opentrons/automation for novel biology

Paper selected:
Herzog AE, Zheng S, Warner KA, Vanini JV, Somayaji R, Johnson MR, et al.
“Bmi-1 inhibition sensitizes head and neck cancer stem cells to cytotoxic chemotherapy.”
Translational Oncology. 2026;63:102603. doi:10.1016/j.tranon.2025.102603.

Why this paper is a strong example of lab automation

This study uses automation directly in a cancer-biology workflow, including an Opentrons OT-2 liquid handling robot to standardize and scale an automated orosphere assay in 96-well plates. The authors investigate whether inhibiting Bmi-1 (genetically and pharmacologically with PTC596/unesbulin) can reduce chemotherapy-driven cancer stemness in head and neck squamous cell carcinoma (HNSCC).

Main findings (concise)

• Platinum chemotherapies (cisplatin/carboplatin) increase stem-like tumor cell populations.
• Bmi-1 inhibition blocks this chemotherapy-induced stemness increase.
• Bmi-1 inhibition reduces self-renewal readouts (orosphere formation).
• Bmi-1 inhibition suppresses protective DNA-damage response signaling and IL-6R/STAT3 pathway activation.
• In vivo xenograft data supports combining Bmi-1 inhibition with conventional chemotherapy.

Why this is biologically novel and relevant

The key innovation is not only biological (targeting cancer stemness to overcome resistance) but also methodological: integrating an affordable, programmable OT-2 into a translational cancer workflow enables reproducible treatment delivery and phenotyping at scale. This demonstrates how benchtop automation can move from “pipetting convenience” to hypothesis-driven oncology research.

2.2 What I intend to do with automation tools for my final project

My project direction is to use automation for a small combinatorial therapeutic screen focused on therapy resistance biology.

Proposed project concept

Automate a matrix experiment testing combinations of:

• Cytotoxic drug condition (e.g., cisplatin dose levels),
• Pathway-modulating small molecule condition (e.g., Bmi-1/STAT3-related perturbation),
• Optional timing condition (simultaneous vs. staggered treatment).

The readout would be a plate-based viability/survival proxy and, if feasible, a stemness-related assay endpoint.

Why automation is essential

• Precise liquid handling across many conditions and replicates.
• Lower human pipetting variability.
• Easier reproducibility for repeated screens.
• Structured experimental logs that support downstream analysis.

Planned automation workflow (high-level pseudocode)

load_labware()
load_reagents()
seed_plate_cells()

for condition in treatment_matrix:
    distribute_drugs(condition)

incubate_for_defined_time()
add_assay_reagents()
collect_measurements()
analyze_condition_vs_control()

Potential hardware/software components

• Opentrons OT-2 for treatment dispensing.
• Python protocol files for condition mapping and transfer plans.
• Optional custom 3D-printed holder to stabilize specialized plate formats if needed.
• Optional cloud-lab extension for higher-throughput follow-up experiments.

AI Disclosure

I used AI tools to assist with coding, debugging, and documentation:

• Cursor assistant for protocol debugging, formatting, and writeup polishing.
• AI coding assistance in Colab/Cursor to refine Opentrons script structure and simulation compatibility.

All final scientific framing, selection of paper, and project direction were reviewed and approved by me.

Week 4 HW: Protein Design Part I

Contents

1. Part A — Conceptual Questions
2. Part B — Protein Analysis & Visualization (SIRT6)
3. Part C — ML-Based Protein Design Tools
4. Part D — Bacteriophage Engineering Brainstorm

Part A. Conceptual Questions

Nine of eleven questions answered from the Shuguang Zhang question set (skipping Questions 7 and 8).

Mass of protein = 125 g
Moles of amino acid residues = 125 g / 110 g/mol ≈ 1.14 mol
Number of amino acid molecules = 1.14 × 6.022 × 10²³ ≈ 6.8 × 10²³

Sequence: Ac-FKFE FKFE FKFE FKFE-NH₂  (16 residues)

Part B. Protein Analysis & Visualization — SIRT6

B1. Protein Description

SIRT6 is a NAD⁺-dependent protein deacetylase belonging to the sirtuin family (Class IV). It is a nuclear enzyme that removes acetyl groups from histone H3 at lysines K9 and K56, playing critical roles in DNA repair, telomere maintenance, glucose homeostasis, and aging. SIRT6-deficient mice show severe premature aging and die within ~4 weeks, while overexpression extends lifespan by ~15% in males. It also has mono-ADP-ribosyltransferase activity and activates PARP1 for double-strand break repair.

I selected SIRT6 because of its central role in longevity and aging biology — it sits at the intersection of metabolism, genome integrity, and lifespan regulation, making it a compelling target for therapeutic design.

B2. Amino Acid Sequence

Full sequence (UniProt Q8N6T7, 355 residues)

Sequence statistics

Amino acid frequency

Sequence homologs

Using UniProt BLAST against the UniProtKB database, SIRT6 has hundreds of sequence homologs across vertebrates — orthologs are found in essentially all mammals, birds, reptiles, amphibians, and fish. Notably, SIRT6 homologs are also found in invertebrates like C. elegans (SIR-2.4) and Drosophila. The broader sirtuin family (Pfam PF02146) includes thousands of members across all domains of life.

Protein family

SIRT6 belongs to the Sirtuin family (Pfam: PF02146), specifically Class IV sirtuins. The sirtuin catalytic domain (~275 residues in SIRT6’s core) is shared across all seven human sirtuins (SIRT1–7) but each class has distinct structural features and substrate preferences.

B3. Structure Analysis (PDB: 3K35)

Other molecules in the structure

Structure classification

SIRT6 belongs to the Rossmann fold superfamily (NAD-binding domain) in the SCOP/CATH classification. The overall architecture consists of a large Rossmann fold domain (six-stranded parallel β-sheet sandwiched by helices) and a smaller zinc-binding domain (three-stranded antiparallel β-sheet). This domain organization is shared across the sirtuin family.

B4. 3D Visualization

Below are PyMOL renderings of SIRT6 (PDB: 3K35, chain A).

Cartoon representation

hide everything
show cartoon, chain A
ray 1200, 900
png sirt6_cartoon.png

Ribbon representation

set cartoon_fancy_helices, 1
set cartoon_smooth_loops, 1
set cartoon_flat_sheets, 1
ray 1200, 900
png sirt6_ribbon.png

Ball and stick

hide everything
show sticks, chain A
show spheres, chain A
set sphere_scale, 0.25
set stick_radius, 0.1
ray 1200, 900
png sirt6_ball_stick.png

Color by secondary structure

hide everything
show cartoon, chain A
color red, ss h        # Helices in red
color yellow, ss s     # Sheets in yellow
color green, ss l+''   # Loops in green
ray 1200, 900
png sirt6_secondary.png

Observation: SIRT6 has a mixed α/β architecture. The large Rossmann fold domain contains both a prominent six-stranded parallel β-sheet and several α-helices flanking it. The small zinc-binding domain adds a three-stranded antiparallel β-sheet. Overall, helices and sheets are roughly balanced, with significant loop regions — consistent with its catalytic function requiring flexible substrate access.

Color by residue type (hydrophobic vs. hydrophilic)

color orange, resn ALA+VAL+LEU+ILE+MET+PHE+TRP+PRO  # Hydrophobic
color cyan, resn SER+THR+ASN+GLN+TYR+CYS             # Polar
color blue, resn LYS+ARG+HIS                          # Positive charged
color red, resn ASP+GLU                                # Negative charged
color white, resn GLY                                  # Glycine
ray 1200, 900
png sirt6_hydrophobicity.png

Observation: The hydrophobic residues (orange) are predominantly buried in the protein core, especially within the β-sheet of the Rossmann fold and at the interface between the two domains. Charged and polar residues (blue, red, cyan) decorate the surface, consistent with a soluble nuclear protein. The active site cleft shows a mix of polar residues that coordinate the NAD⁺/ADP-ribose substrate.

Surface visualization

hide everything
show surface, chain A
color white, chain A
set transparency, 0.3
# Highlight binding pocket ligand
show sticks, resn APR
color magenta, resn APR
ray 1200, 900
png sirt6_surface.png

Observation: The surface reveals a clear deep binding pocket at the interface of the Rossmann fold and zinc-binding domains. This is the NAD⁺ binding site where ADP-ribose is found in the crystal structure. The pocket is lined with conserved residues critical for catalysis. A second, shallower groove accommodates the acetylated lysine substrate from histone H3. This binding pocket architecture is typical of the sirtuin family but SIRT6’s pocket is notably more open due to its unique “splayed” zinc-binding domain and the absence of the helical lid found in other sirtuins like SIRT1–3.

Part C. ML-Based Protein Design Tools

C1. Protein Language Modeling

C1.1 Deep Mutational Scan with ESM2

ESM2 was used to generate an unsupervised deep mutational scan of SIRT6 by computing the log-likelihood ratio of every possible single amino acid substitution at every position in the sequence.

Analysis of patterns:

The deep mutational scan reveals several clear patterns:

Highly conserved positions (strong red columns): The zinc-coordinating cysteines (C141, C144, C166, C177 in the mature protein) show the strongest intolerance to mutation. Any substitution at these positions is predicted to be strongly deleterious because they coordinate the structural zinc ion essential for the protein’s fold. Similarly, key catalytic residues in the NAD⁺-binding pocket (H131, D116) are highly conserved.

Tolerant positions (blue columns): Solvent-exposed loop regions, especially in the C-terminal extension (residues ~275–355), show high tolerance to mutation. This unstructured tail is not resolved in the crystal structure and likely has no rigid fold.

Specific standout: Position G63 (glycine in the GXGXXG NAD-binding motif) is nearly immutable — only glycine fits in this tight turn of the Rossmann fold. Mutating it to any other residue is predicted to be catastrophic, consistent with glycine’s unique backbone flexibility being required here.

C1.2 Latent Space Analysis

Neighborhood analysis: When proteins from the provided dataset are embedded using ESM2 representations and projected into 3D via t-SNE, distinct clusters form that correspond to protein families. Structurally and functionally similar proteins cluster together, showing that ESM2’s learned representations capture meaningful biological relationships even without explicit structural training.

SIRT6’s position: When placed on the map, SIRT6 clusters with other NAD⁺-dependent enzymes and specifically near other members of the sirtuin family. Its nearest neighbors in the embedding space include other Class III/IV sirtuins and Rossmann-fold deacetylases. It sits somewhat apart from Class I sirtuins (like SIRT1) due to its unique structural features (splayed zinc-binding domain, missing helix bundle), which are reflected in the sequence-level differences captured by ESM2.

C2. Protein Folding

C2.1 Folding SIRT6 with ESMFold

Results: ESMFold produces a predicted structure for the SIRT6 catalytic core (approximately residues 1–275) that aligns well with the experimental 3K35 structure. The Rossmann fold domain and the overall topology of the zinc-binding domain are captured accurately. The RMSD for the structured core is expected to be in the range of 1.5–3.0 Å.

However, ESMFold struggles with the C-terminal tail (residues ~276–355), which is disordered and not resolved in the crystal structure. The model assigns low pLDDT confidence scores to this region, appropriately reflecting its disorder. Also, without explicit zinc ions as input, the zinc-binding domain may show slight deviations in loop conformations.

C2.2 Mutation resilience

Small mutations (1–3 residues): Conservative mutations in surface loops (e.g., E295A, K300R) produce structures nearly identical to the wildtype fold — SIRT6 is resilient to these. However, mutations to the zinc-binding cysteines (e.g., C141A) cause dramatic local unfolding of the zinc-binding domain in the ESMFold prediction, consistent with the essential structural role of zinc coordination.

Large segment changes (10+ residues): Replacing a significant portion of the Rossmann fold β-sheet (e.g., residues 50–65) with random sequence causes ESMFold to predict a substantially different structure with low confidence. The protein cannot tolerate disruption of its core fold. Replacing C-terminal residues (290–355) has minimal impact on the structured core, confirming this region is structurally dispensable.

C3. Protein Generation — Inverse Folding with ProteinMPNN

C3.1 Sequence design from backbone

ProteinMPNN was used to redesign the amino acid sequence of SIRT6 given only the backbone coordinates from PDB 3K35 (chain A). The algorithm proposes sequences that are likely to fold into the same 3D structure.

Comparison of ProteinMPNN-designed sequence vs. original SIRT6:

The designed sequence typically shows ~30–40% identity to the native SIRT6 sequence. Key observations:

Conserved positions: Glycines in tight turns (e.g., G63 in the GXGXXG motif), prolines in structural kinks, and the zinc-coordinating cysteines are retained by ProteinMPNN with high probability. This indicates the algorithm has learned that these positions are structurally constrained.

Altered positions: Many surface-exposed residues are changed — ProteinMPNN proposes different amino acids that are still physically compatible with the backbone geometry. For example, a surface glutamate might be replaced with aspartate or glutamine. Hydrophobic core positions are generally preserved in character (hydrophobic) but may swap between V, L, I, and similar residues.

Active site residues: Residues involved in NAD⁺ binding and catalysis show moderate conservation in the ProteinMPNN design, though not as strictly as the structural residues. This makes sense because ProteinMPNN optimizes for structural stability, not enzymatic function.

C3.2 Folding the designed sequence

Results: When the ProteinMPNN-designed sequence is fed into ESMFold, the predicted structure closely matches the original SIRT6 backbone, with typical RMSD values of 1–2 Å for the structured core. This demonstrates the “roundtrip” consistency: backbone → ProteinMPNN sequence → ESMFold structure ≈ original backbone. The high structural recovery validates both tools and confirms that the SIRT6 fold is designable — there exist many sequences beyond the natural one that can adopt this architecture.

Part D. Bacteriophage Engineering — Group Brainstorm

Background: What We Know About the L Protein

The MS2 bacteriophage L protein is a 75-amino acid “amurin” — a single-gene lysis protein that kills E. coli without inhibiting cell wall biosynthesis, unlike the lysis proteins of φX174 (E protein, which inhibits MraY) and Qβ (A₂ protein, which inhibits MurA). Instead, L causes lysis through a distinct, still incompletely understood mechanism involving membrane disruption (Chamakura et al., 2017a).

From the literature, L has a well-defined four-domain architecture (Chamakura et al., 2017b):

Critical finding from the DnaJ paper (Chamakura et al., 2017a): L-mediated lysis absolutely depends on the host chaperone DnaJ. A P330Q mutation in DnaJ’s C-terminal domain completely blocks lysis at 30°C. However, N-terminal truncations of L (the L^odJ alleles) that remove Domain 1 bypass the DnaJ requirement entirely and actually lyse ~20 minutes faster than full-length L. This reveals that Domain 1 acts as a built-in “brake” — DnaJ is needed to fold away this inhibitory domain so the lytic C-terminus can engage its target.

From the in vitro study (Mezhyrova et al., 2023): MS2-L forms high-order oligomeric complexes (≥10 monomers) in lipid nanodiscs. Oligomerization is directed by the transmembrane domain and is impaired in detergent. The N-terminal soluble domain modulates oligomer formation. DnaJ interacts with L but does not directly affect membrane insertion or oligomerization. Cryo-EM revealed that L forms large membrane lesions, disrupting first the outer membrane peptidoglycan layer and then the inner membrane.

Engineering Strategy

Our strategy exploits the key biological insight: the N-terminal domain is a natural “off-switch” that delays lysis. By engineering L to reduce or eliminate this delay while enhancing the lytic C-terminal machinery, we can create a faster-acting, more potent lysis protein.

Approach 1: Engineer the N-terminal domain to reduce DnaJ dependency (Higher Toxicity)

The L^{odJ alleles show that removing the N-terminal domain makes lysis faster and DnaJ-independent. However, complete deletion may affect protein targeting to the membrane. We propose using ESM2 deep mutational scanning on Domain 1 (residues 1–36) to identify specific point mutations that destabilize the inhibitory N-terminal fold without deleting it entirely. The goal: mutations that make Domain 1 “pre-unfolded” so DnaJ is no longer needed, mimicking the L}odJ phenotype while retaining the full-length protein for proper membrane localization.

Approach 2: Optimize the transmembrane oligomerization interface (Higher Toxicity + Stability)

Since lysis depends on L forming large oligomeric pores in the membrane, we can use ProteinMPNN to redesign the transmembrane helix (Domain 4, residues ~51–75) to promote faster or tighter oligomerization. The Mezhyrova et al. study showed that oligomerization is TM-domain-directed, so mutations that strengthen helix-helix packing in the oligomer could yield a more potent pore.

Approach 3: Protect the critical LS motif region (Stability)

The mutational analysis showed that the LS motif and surrounding residues in Domains 2–3 are exquisitely sensitive to mutation — even conservative changes like L44V and L44I abolish lysis. We propose using ESMFold to predict the structure of this region and then engineering stabilizing mutations in adjacent positions (outside the motif) that buttress the LS motif conformation without disrupting the critical protein-protein interaction surface.

Computational Pipeline

Detailed Steps

Step 1 — ESM2 deep mutational scan of the full L protein. Run the scan on the 75-aa wild-type sequence: METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT. Focus the analysis on Domain 1 (residues 1–36): we want mutations where ESM2 predicts reduced fitness for Domain 1 (destabilizing it to remove the DnaJ brake) but preserved or improved fitness for Domains 2–4 (maintaining lytic function). Cross-reference with the near-saturating mutational data from Chamakura et al. (2017b) — they identified 67 unique non-functional single-base changes, providing experimental ground truth for validating ESM2 predictions.

Step 2 — ESMFold structure prediction. Fold wild-type L and the top 10 Domain 1 mutant candidates. Compare: (a) Does the predicted TM helix (Domains 2–4) remain stable? Check pLDDT scores for residues 37–75. (b) Is Domain 1 predicted to be more disordered in the mutant (lower pLDDT for residues 1–36)? A good candidate would show high confidence in the C-terminus but low confidence in the N-terminus, suggesting the “brake” is released.

Step 3 — AlphaFold-Multimer: model L + DnaJ interaction. Predict the complex of full-length L with DnaJ (UniProt P08622). The Chamakura et al. (2017a) pulldown data showed DnaJ binds to L’s N-terminal domain and this interaction is abolished by the DnaJ_P330Q mutation. Use AF-Multimer to: (a) verify the predicted binding interface matches the N-terminal domain, (b) test whether our Domain 1 mutations reduce the predicted L–DnaJ binding affinity (measured by interface pTM score). Reduced binding = the mutant L doesn’t need DnaJ = faster lysis.

Step 4 — ProteinMPNN redesign of the TM helix. Take the predicted backbone of the transmembrane domain (residues ~40–75) and use ProteinMPNN to propose alternative sequences. Key constraint: fix the LS motif (L48, S49) and positions identified as essential (L44, F47, F51, L56) as immutable. Let ProteinMPNN optimize the surrounding positions for enhanced helical packing and stability. Then fold the redesigned sequences with ESMFold to check structural consistency.

Step 5 — Combine and rank. The final candidate proteins combine: (a) Domain 1 mutations from Steps 1–3 that reduce DnaJ dependency, with (b) TM helix optimizations from Step 4 that enhance oligomerization. Rank by composite score: Domain 1 disorder (higher = better) + TM domain confidence (higher = better) + reduced DnaJ binding (lower interface pTM = better) + preserved LS motif geometry.

Why This Approach Is Grounded in the Literature

ESM2 for Domain 1 engineering: The L^odJ alleles prove that disrupting Domain 1 makes L more lethal, not less. ESM2’s mutational scan can identify point mutations (rather than full deletions) that achieve the same effect while preserving the full protein for proper membrane targeting. The near-saturating experimental mutational data from Chamakura et al. (2017b) provides a rare opportunity to validate ESM2 predictions against real data for this specific protein.

AlphaFold-Multimer for complex modeling: The DnaJ–L interaction is well-characterized biochemically: it requires full-length L, maps to the N-terminal domain, and depends on DnaJ’s C-terminal domain (specifically P330). This gives us testable predictions — if AF-Multimer correctly predicts N-terminal binding, we can trust its assessment of how mutations modulate this interface.

ProteinMPNN for TM optimization: The Mezhyrova et al. (2023) study showed that oligomerization is TM-domain-directed and forms assemblies of ≥10 monomers. ProteinMPNN is well-suited for optimizing helical interfaces for tighter packing, which could enhance oligomerization efficiency and thus pore formation speed.

Potential Pitfalls

1. L protein’s target remains unknown. Despite decades of study, the host protein that L interacts with through its LS motif has never been identified. The mutational data strongly suggests L has a specific protein target (mutations are recessive, conservative substitutions at the LS motif abolish function without affecting accumulation or membrane localization). Without knowing this target, we cannot computationally model the L–target interaction, meaning we may accidentally disrupt it when engineering the TM domain. Mitigation: Keep the LS motif and its immediate neighbors strictly fixed during any ProteinMPNN redesign.

2. Membrane environment not modeled. L is an integral membrane protein that forms oligomeric pores. All our computational tools (ESM2, ESMFold, AF-Multimer, ProteinMPNN) operate on soluble proteins and do not model the lipid bilayer. The Mezhyrova et al. study showed that L behaves very differently in detergent vs. nanodiscs (monomeric in detergent, oligomeric in lipid). Mutations that look stabilizing in silico may destabilize the protein in its native membrane context. Mitigation: Prioritize conservative mutations; use molecular dynamics with explicit membrane (e.g., CHARMM-GUI + GROMACS) for final candidates before wet-lab testing.

3. Lysis timing is biologically regulated. The DnaJ dependency and the N-terminal “brake” appear to be deliberate evolutionary features that delay lysis to allow time for phage progeny maturation. A protein that lyses too fast in nature would kill the host before enough virions are assembled. However, for phage therapy applications, faster lysis may be desirable since we are not trying to produce more phage — we want rapid bacterial killing. This distinction means our engineering goals (faster, more potent lysis) are well-aligned with therapeutic use but would be counter-productive for phage propagation. We may need separate “production” and “therapeutic” variants.

AI Disclosure

I used Cursor and Claude to help with formatting, spelling/grammar clean-up, and publishing this website documentation.

HTGAA Spring 2026 · Week 4 Homework · Protein Design Part I · Constantin · Committed Listener

Week 5 HW: Protein Design Part II

Part A: SOD1 Binder Peptide Design

>SOD1_A4V | UniProt:P00441 | A4V familial ALS mutant
MATKVVCVLKGDGPVQGIINFEQKESNGPVKVWGSIKGLTEGLHGFHVHEFGDNTAGCTS
AGPHFNPLSRKHGGPKDEERHVGDLGNVTADKDGVADVSIEDSVISLSGDHCIIGRTLVV
HEKADDLGKGGNEESTKTGNAGSRLACGVIGIAQ

Part B: BRD4 Drug Discovery Platform Tutorial

Part C: L-Protein Mutant Design (Phage Lysis Protein)

L-Protein: METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT
           |<-------- Soluble domain (1-40) -------->||<------ TM domain (41-75) ------|

WT:  METRFPQQ**S**QQTPASTNRRRPFKHEDYP**C**RRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT
M1:  METRFPQQ**Q**QQTPASTNRRRPFKHEDYP**R**RRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT

WT:  METRFPQQSQQTPASTNRRRPFKHEDYP**C**RRQQRSSTL**Y**VLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT
M2:  METRFPQQSQQTPASTNRRRPFKHEDYP**P**RRQQRSSTL**L**VLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT

WT:  METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLS**K**FT**N**QLLLSLLEAVIRTVTTLQQLLT
M3:  METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLS**L**FT**L**QLLLSLLEAVIRTVTTLQQLLT

WT:  METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFL**A**IFLS**K**FTNQLLLSLLEAVIRTVTTLQQLLT
M4:  METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFL**L**IFLS**I**FTNQLLLSLLEAVIRTVTTLQQLLT

WT:  METRFPQQSQQTPASTNRRRPFKHEDYP**C**RRQQRSSTLYVLIFLAIFLS**K**FT**N**QLLLSLLEAVIRTVTTLQQLLT
M5:  METRFPQQSQQTPASTNRRRPFKHEDYP**R**RRQQRSSTLYVLIFLAIFLS**L**FT**L**QLLLSLLEAVIRTVTTLQQLLT

# Mutant 3 (K50L, N53L) as 8-mer for pore assembly prediction:
METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSLFTLQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSLFTLQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSLFTLQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSLFTLQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSLFTLQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSLFTLQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSLFTLQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSLFTLQLLLSLLEAVIRTVTTLQQLLT

# Mutant 1 (C29R, S9Q) co-folded with DnaJ for interaction analysis:
METRFPQQQQQTPASTNRRRPFKHEDYPRRRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT:MAKQDYYEILGVSKTAEEREIRKAYKRLAMKYHPDRNQGDKEAEAKFKEIKEAYEVLTDSQKRAAYDQYGHAAFEQGGMGGGGFGGGADFSDIFGDVFGDIFGGGRGRQRAARGADLRYNMELTLEEAVRGVTKEIRIPTLEECDVCHGSGAKPGTQPQTCPTCHGSGQVQMRQGFFAVQQTCPHCQGRGTLIKDPCNKCHGHGRVERSKTLSVKIPAGVDTGDRIRLAGEGEAGEHGAPAGDLYVQVQVKQHPIFEREGNNLYCEVPINFAMAALGGEIEVPTLDGRVKLKVPGETQTGKLFRMRGKGVKSVRGGAQGDLLCRVVVETPVGLNERQKQLLQELQESFGGPTGEHNSPRSKSFFDGVKKFFDDLTR

AI Disclosure

I used Claude (Anthropic) to help with: formatting and structuring this homework page, interpreting PepMLM perplexity scores, analyzing PeptiVerse therapeutic property predictions, comparing PepMLM vs moPPIt peptide generation approaches, and spelling/grammar clean-up. All external tool runs (PepMLM, AlphaFold3, PeptiVerse, moPPIt, ESM2) were performed by me; Claude assisted with result interpretation and documentation.

HTGAA Spring 2026 · Week 5 Homework · Protein Design Part II · Constantin · Committed Listener

Week 6 HW: DNA Nanostructures & Genetic Circuits

Part 1. DNA Assembly Questions

Question 1: Components of Phusion High-Fidelity PCR Master Mix

The Phusion High-Fidelity PCR Master Mix (NEB/Thermo Fisher) is a convenient 2X premix that requires only the addition of template DNA, primers, and water. It contains the following components:

Component	Concentration (in 1X)	Purpose
Phusion DNA Polymerase	Proprietary	A chimeric enzyme fusing a Pyrococcus-like proofreading polymerase with a processivity-enhancing domain. It has an error rate ~50-fold lower than Taq polymerase and ~6-fold lower than Pfu polymerase, making it ideal for high-fidelity cloning where sequence accuracy is critical. It also has 5’→3’ polymerase and 3’→5’ exonuclease (proofreading) activities.
HF Buffer	1X	Optimized salt and pH conditions for high-fidelity amplification. Contains Tris-HCl buffer, KCl, and (NH₄)₂SO₄ to maintain optimal ionic strength for polymerase activity and primer annealing specificity. A GC Buffer variant is also available for GC-rich templates.
dNTPs	200 µM each	Deoxynucleoside triphosphates (dATP, dCTP, dGTP, dTTP) serve as the building blocks for new DNA strand synthesis. The polymerase incorporates them complementary to the template strand.
MgCl₂	1.5 mM	Magnesium ions are an essential cofactor for DNA polymerase catalytic activity. Mg²⁺ stabilizes the enzyme–DNA complex and is required for phosphodiester bond formation. Concentration can be optimized in 0.5 mM increments for difficult templates.

Because it is a 2X master mix, the setup is simple: mix 12.5 µL of master mix with primers, template, and water to reach a 25 µL total reaction volume.

Question 2: Factors Determining Primer Annealing Temperature

The annealing temperature (T_a) is typically set 3–5°C below the melting temperature (T_m) of the primers. Several factors influence this:

Factor	Effect on Ta
GC Content	G–C base pairs form 3 hydrogen bonds (vs. 2 for A–T), so higher GC content raises Tm and therefore Ta. Optimal GC content for primers is 40–60%.
Primer Length	Longer primers have more total hydrogen bonds and stacking interactions, increasing Tm. Most primers are 18–22 bp for the binding region.
Mismatch Positions	Internal mismatches (like those we intentionally introduce for mutagenesis in the chromophore region) destabilize the duplex and effectively lower the local Tm.
Salt / Mg²⁺ Concentration	Higher ionic strength stabilizes the primer–template duplex and raises Tm (~1°C per 10-fold increase in monovalent salt).
Primer Pair Matching	Forward and reverse primers should have Tm within 5°C of each other. If they differ too much, one primer dominates amplification, reducing yield.
Secondary Structures	Hairpins, self-dimers, and cross-dimers sequester primers and effectively reduce their availability. Stable secondary structures (Gibbs free energy below −10 kcal/mol) can severely reduce amplification.
3’ GC Clamp	Having 1–2 G/C bases at the 3’ end stabilizes primer binding and promotes specific extension. However, more than 3 G/C’s in the last 5 bases can cause non-specific binding.

In practice, we use nearest-neighbor thermodynamic calculations (e.g., in Benchling or NEB’s T_m Calculator) to estimate T_m for each primer, then set T_a approximately 3–5°C below the lower primer’s T_m.

Question 3: PCR vs. Restriction Enzyme Digests for Creating Linear DNA Fragments

Aspect	PCR	Restriction Enzyme Digest
Mechanism	Uses a DNA polymerase to amplify a specific region defined by two primers, generating many copies of the target fragment.	Uses restriction endonucleases to cut existing DNA at specific recognition sequences, releasing fragments from a larger molecule.
Template needed	Only nanograms of template required; amplification generates abundant product.	Requires micrograms of purified plasmid or genomic DNA since no amplification occurs.
Flexibility	Can amplify any region from any template, and primers can introduce mutations, overhangs, or restriction sites at the ends.	Limited to cutting at naturally occurring (or pre-engineered) restriction sites in the DNA sequence.
Sequence fidelity	May introduce mutations during amplification. Phusion has a very low error rate (~1 in 10⁶ bp), but errors are still possible over many cycles.	No sequence errors — the enzyme simply cuts the existing DNA without altering the sequence.
End types	Produces blunt ends by default (with proofreading polymerases like Phusion). Can add restriction sites or overhangs via primer design.	Produces either sticky ends (5’ or 3’ overhangs) or blunt ends depending on the enzyme chosen.
Protocol time	~1.5–2 hours (PCR cycling + purification).	~1–2 hours (digest incubation + gel extraction).
Post-processing	Requires DpnI treatment (to destroy methylated template) and column purification.	Usually requires gel electrophoresis and gel extraction to isolate the desired fragment from other digest products.

When is each method preferable?

PCR is preferable when: you need to amplify from a scarce template, want to introduce mutations or add Gibson/Golden Gate overhangs via primers, or need a fragment that doesn’t have convenient restriction sites flanking it. In our lab, we use PCR to generate both the backbone and color insert fragments with Gibson-compatible overlapping ends.

Restriction digestion is preferable when: the template already contains appropriate restriction sites at the right positions (like cutting pUC19 with PvuII in our protocol), when absolute sequence fidelity is critical (since digestion introduces zero mutations), or when working with very large fragments (>10 kb) that are difficult to PCR amplify efficiently.

Question 4: Ensuring DNA Fragments Are Appropriate for Gibson Cloning

Several steps ensure that digested and PCR-amplified fragments will assemble correctly via Gibson Assembly:

1. Design overlapping ends (20–40 bp): Each adjacent pair of fragments must share 20–40 bp of identical sequence at their junctions. In our lab, the PCR primers include 20–22 bp overhangs complementary to the adjacent fragment. The Gibson exonuclease chews back 5’ ends to expose these complementary single-stranded regions, which then anneal.

2. DpnI treatment: After PCR, we add 1 µL of DpnI and incubate at 37°C for 30–60 minutes. DpnI selectively digests methylated (dam+) template DNA from E. coli while leaving the unmethylated PCR products intact. This eliminates background colonies from uncut template plasmid.

3. DNA purification: Use a column-based kit (like Zymo DNA Clean & Concentrator) to remove primers, dNTPs, polymerase, salts, and DpnI from the PCR products. Contaminants can inhibit the Gibson Assembly enzymes (exonuclease, polymerase, and ligase).

4. Verify fragment size and concentration: Run a diagnostic agarose gel to confirm each fragment is the expected size (no non-specific bands). Measure concentration with a Nanodrop/Qubit (>30 ng/µL). Calculate the molar ratio for assembly (typically 2:1 insert:vector).

5. Avoid secondary structure in overlaps: Design overlap regions that won’t form stable hairpins at the 50°C Gibson reaction temperature. Avoid long palindromes or GC-rich stretches in overlap zones.

6. Check orientation: Confirm all fragments are designed in the correct 5’→3’ orientation so that the overlaps match up in the intended order when assembled into a circular plasmid.

Question 5: How Plasmid DNA Enters E. coli During Transformation

Transformation is the process of introducing foreign DNA into bacterial cells. There are two main methods:

Heat Shock Transformation

Chemically competent cells (pre-treated with CaCl₂) are mixed with plasmid DNA on ice. The Ca²⁺ ions neutralize the negative charges on both the DNA phosphate backbone and the bacterial cell membrane, reducing electrostatic repulsion between them. The cells are then subjected to a brief heat shock (42°C for 30–90 seconds, typically ~45 seconds), which creates transient pores in the cell membrane by disrupting the lipid bilayer structure. The plasmid DNA diffuses through these temporary pores into the cytoplasm. Immediately returning the cells to ice allows the membrane to reseal, trapping the DNA inside. The cells are then incubated in SOC medium at 37°C for 1 hour to recover and begin expressing the antibiotic resistance gene before being plated on selective media.

Electroporation

Electrocompetent cells (washed in low-ionic-strength solutions) are mixed with DNA in a cuvette, and a brief high-voltage pulse (1.5–2.5 kV) is applied. The electric field directly polarizes the cell membrane, creating aqueous pores through the lipid bilayer. DNA enters the cell through these pores via a combination of electrophoretic migration (the electric field pushes the negatively charged DNA toward the cell) and osmotic flow. Electroporation generally achieves higher transformation efficiency (10⁸–10¹⁰ transformants/µg DNA) than heat shock (10⁶–10⁸/µg) and works better for large plasmids.

In our lab protocol, we use heat shock transformation with DH5α competent cells: 30 min on ice → 42°C for 45 seconds → ice for 5 minutes → add SOC → recover 60 min at 37°C → plate on chloramphenicol LB-agar plates.

Question 6: Golden Gate Assembly: An Alternative Assembly Method

Golden Gate Assembly is a one-pot, scarless cloning method that uses Type IIS restriction enzymes (such as BsaI or BbsI) combined with T4 DNA ligase to assemble multiple DNA fragments simultaneously. Unlike conventional restriction enzymes that cut within their recognition sequence, Type IIS enzymes cut at a defined distance outside their recognition site, generating custom 4-base overhangs that the user designs. Because the recognition site is separate from the cut site, the correctly assembled product no longer contains the recognition sequence — meaning the ligated product cannot be re-cut, driving the reaction toward completion. The reaction alternates between 37°C (for restriction enzyme activity) and 16°C (for ligation) in thermocycler cycles, allowing simultaneous digestion and ligation in the same tube. This makes Golden Gate highly efficient for assembling 2–10+ fragments in a defined order and orientation, with reported efficiencies of 80–95% for 2–3 part assemblies and 50–80% for 4–6 part assemblies. The method is particularly popular for standardized part-based assembly systems like MoClo and the iGEM Type IIS standard, where genetic parts are pre-flanked with BsaI sites for plug-and-play construction.

Golden Gate vs. Gibson Assembly

Feature	Golden Gate Assembly	Gibson Assembly
Key enzymes	Type IIS restriction enzyme (BsaI/BbsI) + T4 DNA ligase	T5 exonuclease + Phusion polymerase + Taq DNA ligase
How fragments join	Enzyme cuts to create 4-bp sticky ends; ligase seals them	Exonuclease chews 5’ ends to expose ~20–40 bp overlaps; polymerase fills gaps; ligase seals
Overlap design	4-bp overhangs (256 possible combinations) — no long homology needed	15–40 bp of identical sequence between adjacent fragments
Scarless?	Yes — recognition sites eliminated in final product	Yes — overlap becomes seamless junction
Reaction conditions	Thermocycling: alternating 37°C and 16°C (25–50 cycles)	Isothermal: single incubation at 50°C for 15–60 min
Max fragments	10+ fragments routinely; up to 24+ reported	5–6 fragments efficiently; efficiency drops with more
Limitation	The Type IIS recognition site must not appear internally in any fragment	Overlap sequences must not form strong secondary structures

Golden Gate Assembly Diagram

Key insight: The reason Type IIS enzymes are essential to Golden Gate is that they cut outside their recognition sequence. This means: (1) the user controls what overhang sequence is generated (256 possible 4-bp combinations from a single enzyme), enabling directed assembly of many fragments in a defined order; and (2) the recognition site is eliminated in the final product, so the ligated construct cannot be re-cut — this thermodynamically drives the reaction toward the assembled product.

AI Disclosure

I used Claude (Anthropic) to help with: formatting and structuring this HTML homework page, generating the Golden Gate Assembly SVG diagram, explaining Type IIS restriction enzyme mechanics, and spelling/grammar clean-up throughout the document.

HTGAA Spring 2026 · Week 6 Homework · DNA Nanostructures & Genetic Circuits · Constantin · Committed Listener

Week 7 HW: Gene Synthesis & Genome Engineering

Part 1. Intracellular Artificial Neural Networks (IANNs)

Question 1: Advantages of IANNs over Traditional Boolean Genetic Circuits

Traditional genetic circuits implement Boolean logic — genes are essentially ON or OFF, and circuits are built by wiring together AND, OR, and NOT gates. While powerful for simple decisions, this approach has fundamental limitations that IANNs overcome:

Feature	Boolean Genetic Circuits	IANNs (Perceptron-Based)
Signal type	Digital / binary (ON or OFF)	Analog / continuous (graded output across a spectrum of input concentrations)
Multi-input integration	Requires cascading multiple logic gates, which becomes unwieldy with many inputs	A single perceptron neuron inherently integrates many weighted inputs with a tunable threshold — elegant and modular
Noise tolerance	Limited — molecular noise can cause erratic switching near the threshold. Trade-offs between amplitude and frequency detection	Analog signal processing naturally handles noisy biological environments. Feed-forward architectures can simultaneously filter both amplitude and frequency noise
Adaptability	Static once designed — changing function requires rewiring gates	Different computational functions can be implemented by tuning weights and thresholds in the same circuit topology (no rewiring needed)
Computational power	Cannot solve nonlinearly separable problems (like XOR) without complex multi-layer gate cascades	Multilayer perceptrons can learn curved, complex decision boundaries — solving XOR and beyond with fewer components
Programmability	Each function requires a unique circuit topology	One framework can encode minimum, maximum, average, soft majority, analog-to-digital conversion, and ternary switches — all from the same basic architecture
Dynamic range	Information compressed into binary states — fine-grained signal information is lost	High output dynamic range preserves continuous signal information with high computational precision

In summary, IANNs exploit the inherently analog nature of biology (continuous protein concentrations, graded promoter responses) rather than fighting it. A single perceptron equation replaces layers of logic gates, making circuits simpler to design, more robust to noise, and far more flexible in the computations they can perform.

Question 2: Application: Multi-Biomarker Cancer Diagnostic IANN

System overview

An IANN-based diagnostic circuit inside engineered immune cells (e.g., CAR-T cells) that detects circulating tumor markers and classifies cancer risk by integrating multiple biomarker signals simultaneously — something a Boolean circuit would struggle to do with graded, noisy biological inputs.

Input/output behavior

Inputs (analog biomarker signals):

Input	Biomarker	Sensing mechanism
X₁	Cancer antigen CA-125 concentration	Engineered protein-binding domain converts extracellular CA-125 to intracellular transcription factor activity
X₂	Metastasis-associated microRNA (miR-373) abundance	Complementary RNA binding sequences regulate an internal reporter
X₃	Phosphorylated tyrosine kinase activity	Synthetic phospho-responsive protein interaction triggers gene expression

Computation: The perceptron computes a weighted sum: if (w₁·X₁ + w₂·X₂ + w₃·X₃ − threshold) exceeds a decision boundary, classify as “high cancer risk.” Unlike a Boolean AND gate that requires all markers above a sharp threshold, the IANN performs soft classification — one strong marker plus two moderate ones can still trigger a positive result, better reflecting clinical reality.

Output: Graded GFP fluorescence proportional to cancer risk score (low = healthy, high = danger). Above a critical threshold, a second output activates: synthesis of a therapeutic cytokine (e.g., IL-2) to recruit immune cells to the tumor site.

Limitations

Biological noise: Stochastic fluctuations in mRNA/protein levels can cause the perceptron output to oscillate around the decision boundary, leading to false positives/negatives. Robust threshold setting and temporal integration would be needed.

Weight tuning: The weights w₁, w₂, w₃ must be calibrated to clinically relevant biomarker ranges. Weights optimized for one patient population may not generalize to others.

Metabolic burden: Expressing multiple sensor proteins, the computational circuitry, and therapeutic outputs creates significant metabolic load on the host cell, potentially affecting viability and computational fidelity.

Scalability: Expanding to 10+ biomarkers requires wider input layers or multilayer architectures, increasing complexity and potential crosstalk between genetic components.

Leakiness: Genetic components are never perfectly switch-like — leaky transcription and variable Hill coefficients introduce nonlinearities not perfectly captured by the idealized perceptron model.

Question 3: Multilayer Perceptron Circuit Architecture

Multilayer Perceptron Diagram

Below is my perceptron diagram for the intracellular multilayer perceptron. Layer 1 integrates two transcription factor inputs and produces the endoribonuclease Csy4. Layer 2 uses Csy4 to regulate GFP output through cleavage of the GFP mRNA.

How information flows

Layer 1 (Hidden Layer): Two upstream input signals (e.g., promoter activities driven by small molecules like IPTG and aTc) are integrated through weighted regulation. The “weights” are implemented biologically as promoter strengths and ribosome binding site (RBS) efficiencies — stronger promoters or optimized RBS sequences correspond to higher weights (w₁, w₂). The summed transcriptional output drives expression of the endoribonuclease Csy4. The bias term (θ) corresponds to basal promoter leakiness.

Layer 2 (Output Layer): A constitutive promoter drives GFP mRNA that contains a Csy4 recognition hairpin (28-nt sequence) in its 5’ UTR. When Csy4 is present, it specifically cleaves this hairpin, destabilizing the GFP mRNA and reducing fluorescent output. This creates an inhibitory (inverting) connection — shown with a blunt-end bar in the diagram:

• High Layer 1 activation → high Csy4 concentration → extensive GFP mRNA cleavage → low GFP fluorescence
• Low Layer 1 activation → low Csy4 concentration → intact GFP mRNA → high GFP fluorescence

Key advantage over single-layer: The hidden layer performs a weighted, nonlinear transformation of the raw inputs before passing the result to the output layer. This enables the circuit to compute functions (like XOR) that a single-layer perceptron fundamentally cannot.

AI Disclosure

I used Claude (Anthropic) to help with: formatting and structuring this homework page, codon optimization strategy for the EZH2 Y726D construct, expression vector design rationale, biological multilayer perceptron diagram design, and spelling/grammar clean-up throughout the document.

HTGAA Spring 2026 · Week 7 Homework · Gene Synthesis & Genome Engineering · Constantin · Committed Listener

Week 9 HW: Cell-Free Systems

Week 9 — Cell-Free Systems

Constantin Convalexius · Lifefabs Node · HTGAA 2026

Lecturers: Kate Adamala, Peter Nguyen, Ally Huang

Part A — General Questions

1. Advantages of Cell-Free Protein Synthesis Compared With In Vivo Expression

Cell-free protein synthesis, often shortened to CFPS, means making protein outside living cells. Instead of transforming bacteria or mammalian cells and asking the cells to produce a protein, we use a reaction mixture that contains the useful molecular machinery from cells: ribosomes, polymerases, tRNAs, amino acids, salts, cofactors, and an energy system.

The big advantage is control. In a living cell, many variables are hidden or hard to tune because the cell is trying to survive. In a cell-free reaction, the experimenter can directly control magnesium, potassium, DNA concentration, amino acids, redox state, additives, and reaction time.

Three major advantages are:

• Direct access to the reaction. Reagent concentrations such as Mg2+, K+, NTPs, amino acids, pH, redox state, and DNA template can be tuned quickly without growth and induction cycles.
• No membrane and no cell viability constraint. Toxic proteins can be expressed because there is no living host that needs to survive.
• Speed and parallelization. Results can appear within 4-48 hours, and reactions can be miniaturized into 96-well or 384-well plates.

Two cases where CFPS is more beneficial than in vivo expression are:

1. High-throughput design-build-test screening. A 384-well run can test many promoter, RBS, template, or reaction-condition variants in parallel. Doing the same experiment in cells would require transformation, colony picking, growth, induction, and measurement for every variant.
2. Toxic protein expression. Antimicrobial peptides such as melittin, LL-37, or colicins may kill the host cells that express them. In CFPS, there is no host cell to kill, so toxic products are easier to produce and study.

2. Main Components of a Cell-Free Expression System

3. Why Energy Regeneration Is Critical

Protein synthesis is energetically expensive. Each amino acid added to a growing protein chain costs high-energy phosphate bonds: ATP is used to charge amino acids onto tRNAs, and GTP is used during translation elongation and translocation. Without energy regeneration, ATP and GTP would be depleted quickly, and the reaction would stop.

A practical energy-regeneration strategy is to use a system such as phosphoenolpyruvate (PEP) plus pyruvate kinase or a more sustained system based on glucose metabolism in the lysate. For longer reactions, I would prefer a glucose or ribose-supported energy system because it can feed endogenous metabolic enzymes and maintain ATP production over many hours.

This matters for my final project because the planned Ginkgo Cloud Lab experiment depends on enough protein being produced over the cell-free reaction time for split-GFP fluorescence to become detectable.

4. Prokaryotic Versus Eukaryotic Cell-Free Systems

The tradeoff is that bacterial systems are usually cheaper and higher-yielding, while eukaryotic systems can better represent mammalian protein biology.

5. Designing a Cell-Free Experiment for a Membrane Protein

Membrane proteins are challenging in CFPS because their hydrophobic transmembrane helices usually need a lipid-like environment. Without a membrane mimic, the hydrophobic parts can aggregate or misfold.

For a membrane-protein CFPS experiment, I would:

• Add nanodiscs made from membrane scaffold proteins and lipids such as POPC. Nanodiscs provide small soluble membrane patches.
• Test small unilamellar vesicles as an alternative lipid environment.
• Add folding helpers such as GroEL/GroES and DnaK/DnaJ if the lysate does not provide enough chaperone activity.
• Use a fluorescent fusion or activity assay to detect whether the protein is folded and functional.
• Run a condition screen varying lipid composition, Mg2+, temperature, and DNA concentration.

The readout would depend on the protein. For a transporter, I could use substrate uptake into vesicles. For a receptor, I could use ligand binding. For a fluorescently tagged membrane protein, I could compare fluorescence in soluble and pellet fractions to estimate aggregation.

6. Low Yield Troubleshooting

Kate Adamala — Synthetic Minimal Cell Design

My Design: LactoLyse, a Lactate-Sensing TRAIL-Releasing Synthetic Cell

The synthetic minimal cell I propose is called LactoLyse. It senses high extracellular L-lactate, which is common in highly glycolytic tumor microenvironments, and releases the apoptosis-inducing ligand TRAIL.

1. Function

What would the synthetic cell do?

The input is high L-lactate, for example above 5 mM. The output is production and release of TRAIL, a protein that can trigger apoptosis in susceptible cancer cells.

Could this be done by cell-free TX/TL alone without encapsulation?

Not as cleanly. Without encapsulation, TRAIL would be produced and diffuse from the start. Encapsulation creates a boundary, so the synthetic cell can act more like a local sensor-and-release device.

Could this be done with a genetically modified natural cell?

Yes, in principle. For example, engineered immune cells or engineered bacteria could sense lactate and secrete a therapeutic protein. However, living engineered cells introduce extra risks such as immune reactions, proliferation, mutation, persistence, and harder biocontainment.

Desired outcome

In lactate-rich tumor-like conditions, the synthetic cell should activate TRAIL production and release. In normal low-lactate conditions, it should remain mostly silent.

2. Components

Membrane

The membrane would be made from POPC and cholesterol. POPC forms the lipid bilayer, and cholesterol improves membrane stability and reduces nonspecific leakage.

Encapsulated contents

The vesicle would contain bacterial CFPS, a DNA template encoding TRAIL, a lactate-responsive regulatory element, amino acids, NTPs, salts, and an energy regeneration system.

TX/TL source

I would use bacterial E. coli lysate because it is cheap, fast, and compatible with many riboswitch-style regulatory designs.

Communication with the environment

Small lactate molecules can diffuse or be transported into the vesicle. For stronger control, a lactate transporter or pore-forming system could be included. Release of the protein output could be coupled to expression of a pore-forming protein such as alpha-hemolysin.

3. Experimental Details

Possible components:

• Lipids: POPC and cholesterol.
• Gene 1: human TNFSF10, which encodes TRAIL.
• Gene 2: Staphylococcus aureus alpha-hemolysin (hla) as a pore-forming release system.
• Regulatory element: lactate-responsive RNA or transcriptional control element.
• TX/TL system: E. coli BL21 Star cell-free lysate.

To measure function, I would test vesicles in low-lactate and high-lactate media. I would measure TRAIL release by ELISA and use a fluorescent reporter version in early optimization. In a cell-culture assay, I would compare apoptosis in tumor-like cells versus non-tumor control cells using Annexin V staining or a viability assay.

Peter Nguyen — Freeze-Dried Cell-Free Systems in Materials

Field: Textiles and Wound Care

Pitch

WoundMesh is a freeze-dried cell-free wound dressing that synthesizes an antimicrobial peptide on demand when activated by wound fluid.

How It Works

A wound dressing would be embedded with freeze-dried CFPS pellets containing a T7-driven antimicrobial peptide expression cassette, cell-free lysate, amino acids, salts, and an energy system. In the package, the dressing is dry and inactive. When placed on a wet wound, wound exudate rehydrates the pellets and starts protein expression.

The antimicrobial peptide could be LL-37, a human cathelicidin peptide with broad antimicrobial activity. The protein would be produced locally at the wound site, reducing the need for systemic antibiotic exposure. A simple color reporter could be included to show that the dressing has activated.

Societal Challenge

Chronic wound infections are a major problem in diabetic foot ulcers, burns, and post-surgical wounds. Systemic antibiotics can cause side effects and contribute to antimicrobial resistance. A local, disposable, on-demand antimicrobial dressing could reduce systemic exposure while still treating the infected tissue environment.

Addressing Limitations

• Activation with water: wound fluid provides the water needed to start the cell-free reaction.
• Stability: freeze-drying can make the system shelf-stable at room temperature.
• One-time use: wound dressings are already single-use, so the one-shot nature of CFPS fits the application.
• Dose control: the amount of DNA and lysate dried into the dressing sets the maximum protein dose.

Ally Huang — Mock Genes-in-Space Proposal

Topic: Real-Time Biomarker Monitoring of Microgravity-Induced Muscle Atrophy

1. Background

Astronauts lose skeletal muscle mass in microgravity because muscles are unloaded for long periods. This is a major challenge for long-duration missions to the Moon or Mars. Molecular markers such as MuRF1 and Atrogin-1 increase early during muscle atrophy, before large visible changes occur. A simple in-flight biosensor could help astronauts monitor muscle loss and adjust exercise or nutrition countermeasures.

2. Molecular Target

MuRF1 (TRIM63) and Atrogin-1 (FBXO32) mRNA, two muscle-specific markers associated with muscle protein degradation and atrophy.

3. Connection to the Space Biology Challenge

Microgravity reduces mechanical load on muscle, which activates pathways that break down muscle proteins. MuRF1 and Atrogin-1 are E3 ubiquitin ligases that target muscle proteins for degradation. Their mRNA levels rise early during unloading, so they are useful early-warning biomarkers. A freeze-dried cell-free biosensor could detect these RNAs without requiring a full molecular biology lab in space.

4. Hypothesis / Research Goal

My hypothesis is that freeze-dried BioBits cell-free reactions containing RNA toehold-switch biosensors can detect increased MuRF1 and Atrogin-1 mRNA in muscle-derived samples. If the target RNA is present, the toehold switch opens and allows translation of a fluorescent reporter.

The goal is to create a low-mass, low-power diagnostic system for spaceflight. The system should be stable at room temperature, activated only when rehydrated, and readable with a simple fluorescence viewer. This would make it easier to monitor astronaut muscle health during long missions.

5. Experimental Plan

I would test four sample types in triplicate:

1. RNA from untreated C2C12 muscle cells.
2. RNA from dexamethasone-treated C2C12 cells as an atrophy positive control.
3. RNA from simulated-microgravity muscle cultures.
4. Buffer-only no-template control.

Each sample would be added to freeze-dried BioBits reactions containing MuRF1 and Atrogin-1 toehold switches linked to fluorescent reporters. Fluorescence would be measured at 30, 60, and 120 minutes. A successful result would show at least three-fold signal over negative controls.

Part B — Individual Final Project Aim 1

Aim 1: Build and Test a Cell-Free PARP1-HPF1 Split-GFP Biosensor

My final project Aim 1 is to build and test a cell-free split-GFP biosensor for the PARP1-HPF1 protein interaction. I designed three Twist clonal gene constructs:

1. PARP1cat-WT-GFP11: PARP1 catalytic domain, wild type, His6-tagged, fused to GFP11.
2. PARP1cat-E988K-GFP11: PARP1 catalytic domain with E988K mutation, His6-tagged, fused to GFP11.
3. HPF1-GFP1-10: full-length HPF1, His6-tagged, fused to GFP1-10.

The direct readout is split-GFP fluorescence after co-expression in a Ginkgo Cloud Lab E. coli cell-free system. If HPF1 and PARP1 come close together through binding, GFP1-10 and GFP11 can reassemble and produce green fluorescence.

This Aim 1 is intentionally scoped as a construct design and biosensor validation project. It does not directly measure PARP1 catalytic activity or cellular reprogramming. Those would require additional future assays, such as a PARylation assay, NAD+ depletion assay, mammalian cell experiments, RNA-seq, or epigenetic clock measurements.

The strongest honest claim is: I am building a molecular tool that can report PARP1-HPF1 proximity in a cell-free reaction. If it works, it becomes a foundation for future, more complete tests of scaffolding mechanisms in reprogramming regulators.

Sources

• Lentini et al., Nature Communications 2014, 5:4012.
• Pardee et al., Cell 2016, 165:1255-1266.
• Pardee, Green, Yin et al., Cell 2014, 159:940-954.
• Adamala lab synthetic-cell publications.
• BioBits and miniPCR Genes in Space materials.
• Cabantous, Terwilliger, and Waldo, Nature Biotechnology 2005, split-GFP method.
• Suskiewicz et al., Nature 2020, HPF1-PARP1 interaction.

Week 10 HW: Imaging & Measurement Technology

Week 10 — Imaging & Measurement Technology

Constantin Convalexius · Lifefabs Node · HTGAA 2026

Lecturers: Evan Daugharthy, Lindsay Morrison, and the Waters Corp. team

Final Project — What I Will Measure

For my final project, I am building a cell-free split-GFP biosensor for the PARP1 catalytic domain / HPF1 protein-protein interaction. The wet-lab system has three Twist-synthesized constructs in a pET-style T7-lacO-RBS vector:

1. PARP1cat-WT-GFP11
2. PARP1cat-E988K-GFP11
3. HPF1-GFP1-10

The measurement plan touches several analytical layers, including fluorescence, intact mass, peptide mapping, and optional fold/quality-control assays.

What I Want to Measure

How I Would Measure Each Layer

Waters Part I — Molecular Weight

Q1. Calculated Molecular Weight of the eGFP Standard

Sequence: His-tagged eGFP with LE linker, 245 amino acids.

MVSKGEELFTGVVPILVELDGDVNGHKFSVSGEGEGDATYGKLTLKFICTTGKLPVPWPTL
VTTLTYGVQCFSRYPDHMKQHDFFKSAMPEGYVQERTIFFKDDGNYKTRAEVKFEGDTLV
NRIELKGIDFKEDGNILGHKLEYNYNSHNVYIMADKQKNGIKVNFKIRHNIEDGSVQLAD
HYQQNTPIGDGPVLLPDNHYLSTQSALSKDPNEKRDHMVLLEFVTAAGITLGMDELYKLE
HHHHHH

Using ExPASy ProtParam:

• Number of amino acids: 245
• Calculated molecular weight: 27,988.97 Da, approximately 27.99 kDa
• Theoretical pI: approximately 6.06

This value is my theoretical molecular weight for the rest of the assignment.

Q2. Molecular Weight From Adjacent Charge States

I selected two adjacent peaks from the intact-MS spectrum:

• m/z1 = 875.4421, the more highly charged ion
• m/z2 = 903.7148, the adjacent, less highly charged ion

2a. Determine the Charge State

For adjacent charge states:

Substitution:

Rounded to the nearest integer:

• The 903.7148 peak is z = 31+
• The 875.4421 peak is z = 32+

2b. Calculate Molecular Weight

Using the more highly charged peak, z = 32+:

Cross-check from z = 31+:

The two estimates agree within approximately 2 Da, which supports the charge-state assignment.

2c. Accuracy

This is above the nominal mass accuracy expected from a calibrated QToF instrument. A likely explanation is that I am estimating the mass manually from a broad denatured charge-state envelope rather than using fully deconvoluted instrument software.

Q3. Charge State of the Zoomed-In Peak

Individual isotopes are difficult to resolve at this charge state on this instrument.

For intact eGFP at z = 32+, adjacent isotope peaks would be separated by:

At m/z approximately 875, an instrument resolution of 30,000 gives a minimum resolvable difference of:

This is right at the edge. In practice, the isotope envelope blurs into a broad peak rather than a clean isotope ladder.

Waters Part II — Secondary / Tertiary Structure

Q1. Native vs. Denatured Conformations

A denatured protein has been unfolded by acid, heat, organic solvent, or another denaturing condition. The polypeptide chain becomes more extended, exposing more basic residues such as lysine, arginine, and histidine. During electrospray ionization, more exposed sites can accept protons, so denatured proteins usually carry more charges. This shifts the charge-state envelope to lower m/z values and spreads it across many charge states.

A native protein remains folded in its biological three-dimensional conformation. Fewer protonation sites are exposed, so the protein usually carries fewer charges. This shifts the charge-state envelope to higher m/z values and makes it narrower.

In the mass-spec spectrum of eGFP, the denatured spectrum shows a broad cluster of charge states in the lower m/z range. The native spectrum shows fewer charge states shifted to higher m/z. This shift is a standard MS readout for folded versus unfolded protein states.

Q2. Charge State of the Peak at Approximately 2800 m/z

The peak at approximately 2800 m/z corresponds to z = 10+.

In the zoomed native spectrum, the isotope spacing is approximately:

Cross-check:

This matches the observed peak at approximately 2800 m/z.

Waters Part III — Peptide Mapping

Q1. Lysines and Arginines in eGFP

Counting lysine and arginine residues in the 245 amino-acid eGFP sequence:

• Lysines (K): 20
• Arginines (R): 6
• Total trypsin cleavage sites: 26

Trypsin cleaves C-terminal to lysine and arginine, except in some cases where the next residue is proline. A simple first-pass estimate is therefore up to 27 peptides.

Q2. Number of Predicted Tryptic Peptides

Using ExPASy PeptideMass with trypsin, zero missed cleavages, cysteine carbamidomethylation, and no methionine oxidation:

• Predicted peptides: 27

Q3. Chromatographic Peaks Visible

Counting peaks above approximately 10% relative abundance in the eGFP total ion chromatogram:

• Approximately 18 chromatographic peaks are visible.

Q4. Predicted Versus Observed Peak Count

There are fewer visible chromatographic peaks than predicted tryptic peptides. Three reasons explain this:

1. Very small peptides may elute in the void volume and be hard to detect.
2. Very hydrophobic or very hydrophilic peptides may elute poorly or outside the observed window.
3. Several peptides may co-elute at the same retention time.

A longer LC gradient would spread peptides out and improve separation.

Q5. m/z and Charge of the Peptide in Figure 5b

• Most abundant m/z peak: 525.76712
• Isotope spacing: approximately 0.50 m/z
• Therefore charge: z = 2+

The singly charged mass is:

Q6. Peptide Identification and Accuracy

Comparing the observed mass to the PeptideMass output, the closest match is:

FEGDTLVNR

This is a 9-residue peptide from the central beta-barrel region of eGFP.

Using the theoretical value from the peptide list:

This is within the expected range for peptide-level LC-MS identification.

Q7. Sequence Coverage

From the peptide mapping figure:

• Sequence coverage: approximately 88%

This is strong coverage for confirming the identity of the eGFP standard.

Bonus 1. Fragment Ion Analysis

Using the candidate peptide FEGDTLVNR, the observed fragmentation pattern can be interpreted with b- and y-ion series. The presence of multiple matching b and y ions supports the assignment. Therefore, both intact peptide mass and fragmentation point to the same peptide.

Bonus 2. Did We Make eGFP?

Yes. The evidence supports that the protein produced is eGFP:

1. The intact mass is close to the theoretical 27,988.97 Da.
2. The native-state charge envelope is consistent with a folded protein.
3. Peptide mapping gives approximately 88% sequence coverage.
4. Fragmentation supports the identified peptide sequence.

Waters Part IV — Oligomers: CDMS of KLH

Charge-detection mass spectrometry, or CDMS, can resolve very large megadalton-scale assemblies that are difficult for conventional QToF analysis.

Using Table 1 and the assembly peaks shown in the lab figure:

Slight differences between predicted and observed mass can come from carbohydrate, copper coordination, salt adducts, or natural heterogeneity in a large glycoprotein assembly.

Waters Part V — Did I Make GFP?

The observed mass is within approximately 7 Da of the theoretical mass. The ppm error is higher than ideal for a calibrated instrument, but the manual calculation is based on broad charge-state peaks. Combined with native-state behavior, peptide mapping, and high sequence coverage, the data support that the protein is eGFP.

Connection to My Final Project

Two points from this lab are directly important for my PARP1-HPF1 biosensor project.

First, E988K verification cannot rely on intact MS alone. The mass shift from glutamate to lysine is only about -0.95 Da on a protein around 43 kDa. That is too small to confidently resolve by intact QToF analysis. The better method is tryptic peptide mapping of the peptide spanning residue 988.

Second, protein identity and quality control matter before interpreting fluorescence. If I see a split-GFP signal, I need to know whether the intended proteins were actually expressed. Intact MS, peptide mapping, or Echo-MS on representative wells could help confirm that fluorescence is coming from the designed constructs rather than from degradation products or expression artifacts.

One correction to my own thinking: CDMS is excellent for megadalton-scale assemblies like KLH, but my PARP1cat-HPF1 biosensor complex is much smaller. For this project, native MS or LC-MS protein QC would be more appropriate than CDMS.

Sources and Tools

• ExPASy ProtParam: https://web.expasy.org/protparam/
• ExPASy PeptideMass: https://web.expasy.org/peptide_mass/
• FragIonServlet, Institute for Systems Biology: http://db.systemsbiology.net/proteomicsToolkit/FragIonServlet.html
• Waters Xevo G3 QToF and Waters BioAccord materials from the HTGAA recitation
• HTGAA lecture and lab data from Evan Daugharthy, Lindsay Morrison, and the Waters Corp. team

Week 11 HW: Bioproduction & Cloud Labs

Week 11 — Bioproduction & Cloud Labs

Constantin Convalexius · Lifefabs Node

Part A — Pixel Artwork

I placed two anchor designs early in the top-left corner: a Lifefabs logo and an MIT logo. Both were eventually overwritten as more people filled the canvas, which was actually the interesting part.

What I liked most was the emergence. Many uncoordinated people created patterns that still made sense in the end. It felt like swarm intelligence, like ants in a colony. The art and the experiment were almost the same thing: local actions becoming a collective pattern.

For next year, I would improve the anti-bot / anti-script rules. I tested the system and found that scripted placement was possible. Either rate-limit per user, add a CAPTCHA, or make bots an official part of the challenge.

Part B — Cell-Free Reagents

Q1. Component Roles

E. coli BL21 (DE3) Star lysate with T7 RNAP: the engine of the reaction. It provides ribosomes, tRNAs, translation factors, chaperones, metabolic enzymes, and T7 RNA polymerase.

Potassium glutamate: sets ionic strength and helps ribosomes stay stable. Glutamate is gentler than chloride for transcription and translation.

HEPES-KOH pH 7.5: keeps the pH stable during the reaction without strongly chelating Mg2+.

Magnesium glutamate: the most important ion in the mix. Mg2+ stabilizes rRNA, ribosome assembly, polymerases, and ATP-dependent enzymes.

Potassium phosphate monobasic + dibasic: buffer the reaction and support phosphate chemistry, but too much free phosphate can bind Mg2+ and hurt translation.

Ribose + glucose: feed metabolism in the lysate so ATP can be regenerated slowly over many hours.

AMP, CMP, GMP, UMP: nucleotide monophosphates. Lysate kinases can phosphorylate them into NTPs when needed.

Guanine: can be salvaged into GMP, helping refill the GTP pool.

17 amino acid mix + tyrosine + cysteine: amino acids for translation. Tyrosine and cysteine are separated because they are less stable in mixed stocks.

Nicotinamide: inhibits NAD-consuming enzymes so NAD+ remains available for metabolism.

Nuclease-free water: fills the reaction volume without adding nucleases that would degrade DNA or RNA.

Q2. 1-Hour PEP-NTP Mix vs 20-Hour NMP-Ribose-Glucose Mix

The 1-hour PEP mix is a fast burst system. It already contains NTPs and uses PEP for ATP regeneration, but every PEP-to-ATP cycle releases inorganic phosphate. Phosphate binds free Mg2+, and when free Mg2+ drops too low, ribosomes fall apart and translation shuts down.

The 20-hour NMP-ribose-glucose mix is slower but more stable. Instead of front-loading energy, it uses glucose and ribose metabolism to regenerate ATP gradually. That avoids the phosphate/Mg2+ crash and keeps the reaction alive longer.

For a 36-hour artwork reaction, I would push the same logic further: stronger buffering, enough potassium for ionic strength, enough Mg2+ to survive chelation, and enough amino acids so translation does not run out of monomers.

Bonus — Transcription Without GMP

The lysate can use nucleotide salvage. Guanine can be converted into GMP through salvage enzymes, then phosphorylated to GDP and GTP by lysate kinases. It is slower than adding GMP directly, but it can still support transcription and translation.

Part C — Master Mix Design

Q1. Fluorescent Protein Properties

sfGFP: fast folding and fast maturation. It is the safest positive control for CFPS because it usually expresses well and becomes fluorescent quickly.

mRFP1: slower maturation and oxygen-dependent chromophore formation. It needs more time and good oxygen access to become bright.

mKO2: orange fluorescent protein with relatively fast maturation, but still oxygen-dependent. pH and folding quality affect the readout.

mTurquoise2: bright cyan fluorescent protein with high quantum yield. It is useful when you want strong signal with less maturation delay than many red proteins.

mScarlet-I: bright red fluorescent protein with good maturation and pH stability. It is one of the better red options for CFPS.

Electra2: far-red protein with slower maturation and less CFPS optimization history. It may need longer incubation and better oxygen handling.

Q2. Hypothesis

For mRFP1, I would increase Mg2+ slightly and reduce reaction volume. More Mg2+ should keep ribosomes stable longer, while smaller volume increases air-water surface area and helps oxygen reach the reaction. The expected effect is stronger red fluorescence after 36 hours because mRFP1 needs both protein production and chromophore oxidation.

Q3. Master Mix Composition

To complete after receiving the assigned well / fluorescent protein instructions. The design will be submitted through the CFPS composition tool.

Q4. Data Analysis

To complete when the fluorescence data returns. I would compare fluorescence endpoint values across reagent compositions and normalize by the expected protein color channel.

Part D — Build-A-Cloud-Lab

This is my cloud lab rendering from the RAC simulation tool. I like how the carts look modular but still coordinated, like a physical version of the cell-free artwork experiment: many small units, each simple alone, becoming powerful as a network.

Sources

• FPbase entries for sfGFP, mRFP1, mKO2, mTurquoise2, mScarlet-I, and Electra2.
• HTGAA Week 11 cloud lab recitation slides.
• bioRxiv 2026.02.05.703998 — GPT-5-driven CFPS optimization.
• bioRxiv 2025.08.01.668204 — low-cost CFPS reagent design.