Week 1 HW: Principles and Practices

cover image cover image

Final Project

Frugal Benchtop Bioreactors:

Editing the DNA of an organsim is more accessible than ever. Basic lab equipment and plasmid services like GenScript mean that you can dream up your own sequence and express it in an host for around a hundred US dollars. However to unlock the real world impact of gene edits you usually need to be able to scale the production up. The next step in scaling beyond the shaker flask is a bench-top bioreactor where you figure out how to actively manage and optimize your organisms growth and characterstics. This expense of the benchtop stage makes it less accessible than the edit stage even though in many ways the technology involved is simpler. For example a new benchtop bioreactor typically costs tens of thousands of dollars or more. Even used bioreactors costs thousands of dollars.

Substantial progress has been made toward accessible and open bioreactors with efforts like BIO-SPEC: An open-source bench-top parallle bioreactor system, but the Bill Of Materials for a BIO-SPEC system is still over 2000 Euros. Another great project is the Pioreactor which is available as a kit for only $350 US dollars. However the max volume of a Pioreactor is 40mL and it only has a single tank, which limits the ability to do things like media optimization unless you buy multiple Pioeractors.

The goal of my final project would be able to take advantage of fabrication tools like 3d printing to design and build a simple benchtop bioreactor with a bill of materials of around a hundred US dollars, demonstrate the reactors works by scaling up and optimizing production for at least one simple engineered microbe, and then release the plans and software with an open license. The bioreactor should support at least up to 1L of total culture and have multiple reaction tanks to support environmental optimization. A key aspect of benchtop scale up in monitoring and optimizing both the biomass and level of expression, which can be difficult to monitor cheaply. To prove out the concept and provide a demo project for the system the project will also edit an organism like a microbe to express a colored protein so that monitoring and optimizing is easier using inexpensive technology like digital cameras. Eventually a library of scaffold organisms could be developed that allow uses to insert their edits into a the organism in a way that their protein will be co-expressed with the easy to monitor protein like color to enable inexpensive and cheap scale up.

Governance Policy and Actions

Equity And Automony

In line with the desire for to giving as many people as possible the opportunity to learn how to scale up thier bio-engineering projects the primary policy concern is promoting ongoing equity and autonomy.

Actions

  • The project can release all CAD designs, Documentation and, software is released with open licenses that allow people to use and extend. (Most Effective - 1 )
  • The project can ensure the bill of materials only includes items that are broadly available and/or have open licenses. (Minimally Effective - 3)
  • Fab labs and service manufactures can sign up to print and mail kits of 3d printed parts to people who don’t have access to 3d printers. (Minimally Effective - 3)
  • Schools and DIY labs can incorporate this into their curriculum or sponsor people to build them locally in order to expand the number of people who have experience with the process of scaling up synthetic organisms. (Mininimally Effective - 3)
  • An existing organization, like Neosynbio or Frugal Science Academy which support open and frugal biological tools, can sponsor the project and manage the licenses and copyrights. In addition to providing exposure, if the project is successful this allows the project to exist beyond a single person and prevents license and copyright changes that restrict access. (Most Effective - 1)

Biosafety

While a bench-top reactor does not introduce new biological risks it magnifies the ones that already exist in the bio-engineered organism, so a secondary policy goal needs to make be managing this magnified risk.

Actions

  • The project can provide documentation and training materials can reference existing training on bioethics and biosafety. (Moderately Effective - 2)
  • The project and labs can provide known safe demo projects that demonstrate the principles of scale up and optimization with minimal biological risk. (Minimally Effective - 3)
  • Schools and Labs can extend their existing bioethics and biosafety training to explictly discuss how to manage the magnified risk of scaling up a biological process. (Moderately Effective - 2)
  • Governments and regualtors can extend thier existing regulations and logging practices to cover specfic requirements around scaling up or just regulate edits with the assumption that scale up will happen (Moderately Effective - 2)

Homework Questions

Homework Questions from Professor Jacobson

  1. Nature’s machinery for copying DNA is called polymerase. What is the error rate of polymerase? How does this compare to the length of the human genome. How does biology deal with that discrepancy?

The error rate for polymerase is 1 in 10^6 according to slide #8. Slide #10 indicates the human genome is 3.2 Gbp, so a single copy of the genome is likely to have ~1000 errors which is pretty high especially for fast growing cells that replicate once a day, since errors will accumulate with each replication. Biology deals with this by having many layers of error correction and handling beyond the already excellent ones built into the polymerase. One example is the Lamers et al work on MutS from the slides which correct higher level structural errors in the DNA. Beyond error corrections, there are also mechanisms that cause cells with serious errors to self-destruct or be marked for destruction by other cells. This fail-safe removes cells with serious errors from the population so that don’t replicate more.

  1. How many different ways are there to code (DNA nucleotide code) for an average human protein? In practice what are some of the reasons that all of these different codes don’t work to code for the protein of interest?

Slide 6 indicates that the average human protein is about ~1000 base pairs and ~330 amino acids. While the exact number of alternate codings depends on the specific amino acid sequence of the protein, on average each amino acid has ~3 codons that map to it (64/21), so there ~3^330 alternate codings for an average human protein. In practice all of these encodings may not actually work because at the end of the day different codons are still physically and chemically different from other codons which can create differences in the structure of the DNA and translated RNA as indicated in later slides. For example structural differences in RNA chains can impact ribosome translational efficiency, which means a given DNA chain might code for the same protein but make too much or too little of that protein for the organism to survive.

Homework Questions from Dr. LeProust: [Lecture 2 slides]

  1. What’s the most commonly used method for oligo synthesis currently?

From the Jacobson slides and the timeline in the LeProust slides it appears that while the details of the chemistry, level of automation, and miniaturization has been massively improved over the years most olgio synthesis methods are variants of an open loop chemical synthesis with a protection group .

  1. Why is it difficult to make oligos longer than 200nt via direct synthesis?

Also from the Jacobson slides it the error rate for chemical synthesis is 1 in 100 base pairs. That error rate would make it very hard to go much beyond a couple hundred base pairs without getting an error.

  1. Why can’t you make a 2000bp gene via direct oligo synthesis?

    2000 base pairs is much too large for a 1 in 100 error rate. At 2000 bps, even if you sequence the the different olgios produced and then try fo find and amplify a good olgio you have an almost zero chance of getting a good sequence of that length in the first place.

Homework Questions from George Church

I looked at question #1

What are the 10 essential amino acids in all animals and how does this affect your view of the “Lysine Contingency”?

I had to google “What is the lysine contingency” because I had forgotten that part of Jurassic Park. Looking at Wikipedia - Essential Amino Acids it looks like there are 9 amino acids that animals absolute can’t synthesize and 6 more that can’t always be synthesized in sufficient amounts. I was not able to reconcile this vs the 10 amino acids in the question. Independent of that it is clear that the “Lysine Contigency” makes no sense as a form of bio-safety for animals, since all animals already suffer from the lysine contingency and do just fine getting Lysine from the food they eat. Even knocking out the ability to produce a non-essential amino acid like Alanine would not help containment unless the dinosaur species has such a narrow diet that it couldn’t survive any any modern wild plant or animal matter. In particular carnivorous dinosaurs that are happy to eat modern mammals (the ones you most want to contain) could get the full suite of amino acids from their prey. I used Google and their AI answer to verify that all amino acids survive stomach acid and digestion.

Extra Investigation for GRO:

While reading the George Church slides on Geneticially Recoded Organisms (GRO) and the fact that they are immune to viral infection “Swapped genetic code blocks viral infections and gene transfer”, I was driven to do some additional investigation into the risks using Google Gemini knowing how important viral infection is in microbial control in the wild. Some of the main prompts were:

  • “Isn’t immune to natural viruses a little dangerous since that is the main predator/cause of death of most microbes?”
  • “It seems like if they could escape and dump the auxotrophy gene they might have enough of a survival advantage (there are many slow growth bacteria in nature) that they could climb back up the fitness curve over time.”
  • “what is the pragmatic outcome advantage or technology that justifies this risk”

I need to do more investigation in this area, but my first impression is that the researchers involved are thinking deeply about the bio-safety risks and developing layered countermeasures. However many of those countermeasures are fundamentally dependent on responsible actors in a high trust world and it isn’t clear that in a low trust multi-polar world these safeguards will be sufficient? The same economic forces pushing in this direction (virus free bioreactors) may tempt people to bypass the fitness safe-guards for economic benefit. Once a virus immune GRO population becomes established in the wild it isn’t clear to me how you would eradicate them. Even if they are initially slow growing or not dominant in their niche, being immune to viruses is such an incredible fitness advantage that it seems likely GRO microbes will eventually climb the fitness curve and come to dominate all ecosystems. This might take thousands or even millions of years, but seems like a very bad outcome. I guess we would have to hope that natural virus evolution figures out how to bypass the GRO defense before then?