Week 1 HW: Final Project Ideas


Water Hyacinth Cat Litter as a Biological Health Sensor

1. What do I want to biomanufacture, and why?

I want to develop water-hyacinth cat litter that can act as a passive biological health sensor.

I already work with Eichhornia crassipes (water hyacinth), an invasive aquatic plant, and have explored converting its biomass into compressed pellet litter. My proposed next step is to make the litter biologically functional: when a cat urinates, selected pellets or a sensing component associated with the litter would produce a visible colour change in response to a urinary biomarker.

The long-term idea is not for the litter to diagnose disease. Instead, it could provide an early warning of changes in urine chemistry, prompting the cat’s caregiver to seek veterinary testing when necessary.

The system would have two interconnected parts:

Water hyacinth → dried biomass → processed material → pellets → cat litter

and

Urine → biological sensing reaction → colour change → camera detection → change tracked over time

I am interested in this because it connects three areas I already work with: invasive biomass, material development and animal welfare. It asks whether a low-value invasive plant can become the physical platform for an inexpensive biological sensing system.

2. What living system would I use?

For the first prototype, I would investigate a cell-free biosensing system rather than putting living genetically engineered organisms into cat litter.

Cell-free systems use biological machinery outside intact living cells. The sensing reaction could be designed so that a target molecule in urine triggers production or activation of a visible colour signal.

The simplified system would be:

Input: cat urine containing the target biomarker ↓ Recognition: a biological sensing component recognizes the target ↓ Biological reaction: the cell-free system produces or activates a reporter ↓ Output: a measurable colour change

The sensor could eventually be incorporated into selected indicator pellets or a separate sensing region associated with the water-hyacinth litter.

I would not initially attempt to detect many diseases. During the course I would select one experimentally realistic urinary parameter or biomarker and demonstrate the basic principle:

urine sample → biological recognition → visible signal.

One major assumption is that a suitable sensing mechanism can operate reliably in the chemically complex environment of cat urine and can be stored in a dry form until use.

3. How would I use automation?

I would use automation at two very different scales: laboratory development and eventual manufacturing.

During HTBAA: automate sensor development

A liquid-handling robot such as Opentrons could help test many combinations of sensor conditions systematically.

For example, instead of manually preparing dozens of samples, the robot could dispense different concentrations of the target molecule, sensor components and controls into multiwell plates.

A camera or plate reader could then measure the resulting colour response.

The workflow could become:

prepare samples → dispense reagents → run sensing reactions → image/measure colour → compare results → identify the best formulation

This would allow me to investigate questions such as:

• What concentration can the sensor detect?
• How quickly does the colour appear?
• Does urine interfere with the reaction?
• Does the sensor still work after drying and storage?
• Can the biological sensing system survive incorporation into or alongside the litter?

At manufacturing scale: automate biomass processing and sensing

The larger vision is an automated production chain beginning with the invasive plant itself:

harvest → wash → dewater → dry → shred/grind → pelletize → add sensing component → package

Water hyacinth is extremely wet when harvested—published studies report roughly 90–95% initial moisture, making dewatering and drying one of the major processing challenges. Controlled drying has already been demonstrated as substantially faster than conventional sun drying.

Sensors could measure moisture during drying, while temperature and feed rate could be controlled during pellet production. Water-hyacinth pellet studies also show that moisture and processing temperature strongly affect pellet quality, so these are useful parameters to automate.

Finally, Shoobot, an automated litter box, could automate the user end of the system. After the cat urinates, a camera under controlled lighting could photograph the exposed sensing area, quantify colour change, compare it with previous readings and identify unusual changes over time.

The complete automated loop therefore becomes:

remove invasive plant → manufacture litter → biological sensing → machine-readable colour change → longitudinal health information.

4. How much would I need to make, and how long would it take?

(a) How much?

For a meaningful first scale, I would define the target as:

enough sensing litter for 100 cats for one month.

For a back-of-the-envelope estimate, I will initially assume each cat requires approximately 10 kg of pellet litter per month. This is an assumption that I would need to verify experimentally because actual consumption will depend on pellet durability, urine absorption, Shoobot’s sifting system and how frequently litter is replaced.

Therefore:

100 cats × 10 kg/cat/month = 1,000 kg litter/month

So the manufacturing target is approximately:

1 tonne of functional water-hyacinth litter per month.

Water hyacinth contains a very large amount of water. Published measurements commonly place fresh biomass around 90–95% moisture.

If I make the deliberately simplified assumption that fresh plants contain ~5% usable dry matter, then:

1,000 kg dry biomass ÷ 0.05 ≈ 20,000 kg fresh water hyacinth

So producing roughly one tonne of dry biomass-based litter could require on the order of:

20 tonnes of freshly harvested water hyacinth.

This calculation deliberately ignores processing losses, additives and the mass of the sensing system. A real process would therefore require somewhat more biomass.

The biological component is much smaller. If, for example, only 1% of the litter needs to function as indicator material, then:

1,000 kg × 1% = 10 kg of sensing material/month.

This is one of the questions I would investigate: does every pellet need to contain the biosensor, or can a small fraction of sensing pellets provide sufficient information?

Reducing the amount of biological sensing material could dramatically change the economics and scalability of the system.

(b) How long?

The physical biomass processing and biological manufacturing have very different production rates.

Water hyacinth traditionally requires several days of sun drying. A recent controlled dryer study reported reducing approximately 10 kg of wet water hyacinth to below 10% moisture in about 280 minutes, demonstrating why drying would become an important scaling bottleneck.

At that experimental throughput:

20,000 kg ÷ 10 kg/batch = 2,000 batches

which would obviously be impractical with a single small dryer.

This immediately shows that meaningful-scale manufacturing would require continuous or multiple parallel dewatering and drying systems, rather than simply scaling the laboratory apparatus.

Pelletization could then operate as a continuous mechanical process. Published experiments demonstrate that water hyacinth can be pelletized and that parameters including feedstock moisture and temperature affect pellet density and strength.

The biological sensor would initially be produced separately. During prototyping, automated liquid handling could manufacture and test hundreds of small sensor reactions in parallel. Once the formulation is established, manufacturing would shift from laboratory microplates toward batch production of the biological sensing reagent, followed by deposition onto indicator pellets or another sensing substrate.

My long-term target would therefore be a modular system capable of processing approximately:

20 tonnes wet water hyacinth/month → ~1 tonne functional litter/month → litter for ~100 cats/month.

The purpose of this estimate is not to claim that these numbers are optimized, but to identify where scale becomes difficult. The calculation suggests that harvesting, dewatering and drying tonnes of wet biomass may be a much larger physical scaling problem than producing the biological sensor itself.

5. What don’t I know yet?

1. What should I actually detect?

I need to determine which urinary parameter or biomarker is both biologically meaningful and realistically detectable with a simple cell-free or enzyme-based colour reaction.

The first goal should probably be one reliable signal rather than attempting to diagnose multiple diseases.

2. Can a biological sensor survive as part of cat litter?

This may be the central technical question.

The sensing system must survive manufacturing, drying, storage and exposure to the litter environment, but still activate rapidly when urine reaches it.

It may turn out that incorporating the sensor directly into pellets is impossible. A separate indicator pellet, coating, cartridge or sensing surface inside Shoobot may be more practical.

3. Can I distinguish a meaningful biological signal from normal variation?

Cat urine varies naturally in concentration and composition. Water-hyacinth biomass itself may also introduce colour, absorb liquid or interfere chemically with the sensing reaction.

I therefore need to determine whether the colour response is sufficiently specific, repeatable and measurable to distinguish a real change from ordinary variation.

The project in one line

Can an invasive aquatic plant be transformed into cat litter that not only absorbs waste, but becomes part of a biological sensing system capable of revealing changes in an animal’s health?

Final Project