There's too many big sciency words for me to evaluate with not much time. Would be easier to get a sense of the value with a headline statement summarizing the value proposition / theory of change.
Dony Christie
8 months ago
There's too many big sciency words for me to evaluate with not much time. Would be easier to get a sense of the value with a headline statement summarizing the value proposition / theory of change.
Alyssa Riceman
9 months ago
"Imagine if we can produce 2x the amount of Insulin through better codon optimization. This could lower drug prices and even improve time for production of vaccines in a future pandemic."
Do you have alternative examples of the sorts of drugs this would potentially cheapen production of? My understanding of the market for Insulin in particular is that it's already in fact very cheap to manufacture, and supply is constrained primarily, not by price-to-manufacture, but by regulatory limits on who's allowed to manufacture and sell it.
Ross Rheingans-Yoo
9 months ago
Can you explain what your plans are for outcomes where you raise less than $20k? What would you do on $1k, $3k, $10k, or $15k budgets?
Do you have other sources of funding that you're courting? And, given them, how should we think about the counterfactuality of the marginal dollar here from the Impact Certs round?
(Context: I am considering funding for ≥$1k, but am not decided. I could be more interested if there were a smaller pilot experiment that would derisk the scientific risk.)
Aditya Jain
9 months ago
@_rossry Hi, thanks for considering funding the project! In response to both of your comments, here's a more detailed look at the cost budgeting.
TLDR: the $15k projection is our highest expected cost and was chosen to maximize likelihood of success. In reality, cost may be lower once budget is optimized, but hard to predict ahead of time without expert guidance, and securing funding would make it much easier to get that guidance. If costs are lower than raised amount, project can be easily expanded to have more impact. If raised amount is <15k/amount needed to benchmark 30 genes and push industry adoption with GenScript, there are meaningful goals we can achieve by benchmarking 3-10 genes like publishing in a top tier journal to push research community adoption.
Detailed response:
Here is the benchmark sequence database we had planned to use for testing: https://github.com/Lattice-Automation/icor-codon-optimization/tree/v1.4/benchmark_sequences/dna These were selected based on application and use in other research papers which would allow for easier comparison between the ICOR approach and others, and so we don't need to reinvent the methodology. Average BP ~1800. We looked at production via GenScript. One important factor to note is that we need to have the full process done externally from Gene Synthesis, not Plasmid Preparation. This includes steps of adding antibiotic resistance (to select out bacteria which successfully took up the plasmid), adding excision sites around the sequence to insert it into the plasmid, etc. Theoretically, we could do this ourselves which would reduce cost. But, this gives the benefit of being robust, standardized, and would save resources in lab materials & support. Taking this into account, we can leave the rest of the settings default. Ideally, I would prefer to add additional checks like ensuring sterility, producing it without animal products, etc but let's leave it at the bare minimum. Also, note that ~1/3 of the sequences are "Complex" so GenScript is unwilling to give a full cost estimate until they are sure we are ordering it.
I have not requested a full quote as I need to verify some parts of the methodology. But with 40 sequences, the preliminary cost is given as $1960 (shipping to Boston, MA). This comes to $50 per sequence. Now, this cost would actually be higher as they have not added in gene synthesis cost, but with student discount, the bulk discount as we are actually ordering 2-3x this, and institutional discounts, I ball-parked that the cost would stay at around $50/sequence. I understand this is not the most detailed analysis for cost and there may be other competitors which can give us a cheaper rate. If we are funded, my plan is to speak with some PIs and mentors with more wet lab experience who can help me figure out how to optimize cost. I apologize for not having the most detailed budget, but rest assured that I will work with experts on this and if it can be done for cheaper, would use the remaining funds to expand the outcomes (eg: test in multiple host organisms which is becoming more relevant for some specialized drugs)
As for your second question. There are definitely meaningful goals we can achieve with reduced funding! Most published papers which seek to show a new method for optimizing protein expression use far fewer benchmark genes - usually 3-10.
(For example: https://www.nature.com/articles/s41586-023-06127-z#Sec10, for COVID mRNA production and you can see many others in recent research with a similar amount of test sequences. Also side-note: they are also initially evaluating their tool with CAI which we did in our publication and they look for this "sweet spot" in CAI with a human-made computational algorithm below peak CAI due to the problem of too few tRNAs mentioned in the proposal. This is a cool approach, but I suspected our AI approach outperforms this, and this paper was in Nature (!) in 2023 so it's very state of the art).
So with $5k we could test approximately 10 benchmark genes, $3k would still let us do 6, etc. This would result in a paper publication that I expect would be high impact for the research community, would solve the main criticism of ICOR which was that it was not wet lab tested, and would be a major stepping point towards adoption. However, it would not meet GenScript's metrics for adoption in their pipeline and I think it would increase the timeline for it to reach industry.
Grateful for your consideration!
Ross Rheingans-Yoo
9 months ago
@adityajain42 My instinct is that, with the effect sizes you're claiming (+200%?), just 3 genes should give clear enough results to make a full 30-gene experiment a "clear yes" for follow-on funding. So considering the marginal value of the last ~$17k (compared to an initial $3k), I think it mostly comes down to faster timeline at the cost of some probability (30%?) of wasting the costs of genes 4 through 30.
(Let me know if you don't think that 3 genes would be >90% to demonstrate a clearly definitive result, conditional on the full panel returning a positive result.)
I'm not sure how much to care about publishing the intermediate result (since the next step would be going to the full 30-gene panel), so the main delay is in how much longer the two-stage (3+27) experiment would take than a one-stage (30 in parallel) experiment would. (Assuming you don't stop to publish the intermediate result.) Can you give a sense of the additional calendar time for two-stage?
Aditya Jain
9 months ago
I think you're right, I agree with your analysis. I would just note that the way we are predicting 2x increase in output is because we demonstrated a ~27% increase in CAI and an established analysis had shown the correlation between CAI and output, though the strength of that was about 0.65 R^2. So, there is some uncertainty.
As for publication, I had not considered the idea of follow-up funding conditional on the initial 3, which is why I thought that would be a good endpoint, but if there is the possibility of follow-up to reach the 30 sequences needed, then I agree that intermediate publication would not be needed.
So with these factors, I think the additional calendar time for the 2 stage experiment would be approximately 2 months: 1 month to design and receive the plasmids, 1 month to do the experiment. As I mentioned in the initial proposal, I am a student and have other responsibilities for my education, so there may be some additional slowdown, but I do not expect a significant slowdown based on my current schedule.
Ross Rheingans-Yoo
9 months ago
It looks like the primary (65%+) portion of the cost is for plasmid orders at $50 / plasmid. Can you post your diligence on the available suppliers, why you've picked the one you did, and check that this is the wholesale / bulk cost, not the retail / one-off? It seems surprisingly high to me, but I don't have an experience in this niche of the supply chain, so I'd like more info there.
Ross Rheingans-Yoo
9 months ago
typical preliminary wet lab tests involve 30 genes. We would need to order the original plasmid, the modified ICOR plasmid, and the modified industry competitor plasmid. To do robust statistical analysis, we would manufacture 3x. As the cost for each plasmid is ~$50 (depends on size of plasmid sequence), this comes out to about $13,500.
30 \* 3 \* $50 = $4,500 but you've written a total of $13,500. Can you explain the discrepancy?
Ross Rheingans-Yoo
9 months ago
Oh, there are three kinds of plasmid, and it's three copies of each.
I don't know how to delete / edit my prior comment, but feel free to ignore it.
Aditya Jain
9 months ago
No worries, yep as noted "the original plasmid, the modified ICOR plasmid, and the modified industry competitor plasmid." so 30*3*50*3