3 ms·

23andme has made a pretty impressive recovery after the FDA issues a few years ago, although i think the jury is still out on how useful their data will be for

23andme has made a pretty impressive recovery after the FDA issues a few years ago, although i think the jury is still out on how useful their data will be for drug discovery. I think they mostly have genotype data rather than WES, and i think their clinical data is all patient reported? The value of their engaged audience seems clearer, especially if it can help enroll clinical trials faster

randcraw 8y ago
It may be that GSK isn't really interested in the existing 23&me product line, but wants their name recognition and infrastructure. 23 is as close as anyone to delivering a full GWAIS product to the public, which this partnership only enhances. Combined with large capacity cheap deep sequencing by someone like BGI, this partnership would be ideal for GSK to add in-house infrastructure to do a LOT of GWAIS for all kinds of drug target ID efforts or clinical patient stratification. Maybe GSK has decided to target patients very precisely as standard practice in all future drug trials. This makes a lot of sense if you want a new drug to maximize efficacy or minimize toxicity by cherry picking your high responders and low intoxicants, thereby easily avoiding all folks who don't benefit from it, even if no-go patients are in the majority. But it also assumes such cherry labels will be available to physicians and insurers soon, presumably as de rigeur medical practice for the general public. Streamlining a product like 23's might make this initiative fly economically. Done early in a drug trial, like phase II, you might be able to use deep genome data to tune up or rescue a new drug that looks promising but not quite promising enough to continue to phase III. Deep genomic analysis is a very promising way to stratify patients who will benefit from those who won't. This looks to me like GSK has just bought into GWAIS in a big way for thousands or more patients, esp in clinical trials. This partnership can only help both companies, since I don't see a major pharma ever trying to take over 23's existing boutique geek navel-gazer business.
- aaavl2821 8y ago
  Yeah I agree that 23andme name recognition and brand is valuable and maybe best in class for engagement of patients with genomic data. So for clinical trial recruitment and marketing that would help. I'm less convinced of the value of 23 and me data in drug discovery. It could be quite useful but is not best in class. The regeneron genetics center is a much better approach imo -- sequencing data vs genotyping, access to robust medical data, partnerships with leading medical institutions to get patients and data, and a targeted biology driven strategy that reduces the cost of getting a signal. As far as cost per good potential target, 23and me is probably an order of magnitude or more behind Not sure how 23 and me would help GSK analyze genetic subpopulations -- GSK could sequence patients just fine on their own without 23and me and get more robust data than 23 and me genotype data. For biomarker discovery once you have a candidate and are moving into clinical studies there are probably much better targeted solutions than 23 andme Value of 23 and me is 1) brand and 2) data on large number of pts. 2) helps with finding genetic markers with low effect size which isn't really helpful for drug discovery and also finding rare variants with large effect, which is important, but there are better / cheaper ways of doing that (i.e. regeneron)

randcraw 8y ago

It may be that GSK isn't really interested in the existing 23&me product line, but wants their name recognition and infrastructure. 23 is as close as anyone to delivering a full GWAIS product to the public, which this partnership only enhances. Combined with large capacity cheap deep sequencing by someone like BGI, this partnership would be ideal for GSK to add in-house infrastructure to do a LOT of GWAIS for all kinds of drug target ID efforts or clinical patient stratification. Maybe GSK has decided to target patients very precisely as standard practice in all future drug trials. This makes a lot of sense if you want a new drug to maximize efficacy or minimize toxicity by cherry picking your high responders and low intoxicants, thereby easily avoiding all folks who don't benefit from it, even if no-go patients are in the majority. But it also assumes such cherry labels will be available to physicians and insurers soon, presumably as de rigeur medical practice for the general public. Streamlining a product like 23's might make this initiative fly economically. Done early in a drug trial, like phase II, you might be able to use deep genome data to tune up or rescue a new drug that looks promising but not quite promising enough to continue to phase III. Deep genomic analysis is a very promising way to stratify patients who will benefit from those who won't. This looks to me like GSK has just bought into GWAIS in a big way for thousands or more patients, esp in clinical trials. This partnership can only help both companies, since I don't see a major pharma ever trying to take over 23's existing boutique geek navel-gazer business.

aaavl2821 8y ago
Yeah I agree that 23andme name recognition and brand is valuable and maybe best in class for engagement of patients with genomic data. So for clinical trial recruitment and marketing that would help. I'm less convinced of the value of 23 and me data in drug discovery. It could be quite useful but is not best in class. The regeneron genetics center is a much better approach imo -- sequencing data vs genotyping, access to robust medical data, partnerships with leading medical institutions to get patients and data, and a targeted biology driven strategy that reduces the cost of getting a signal. As far as cost per good potential target, 23and me is probably an order of magnitude or more behind Not sure how 23 and me would help GSK analyze genetic subpopulations -- GSK could sequence patients just fine on their own without 23and me and get more robust data than 23 and me genotype data. For biomarker discovery once you have a candidate and are moving into clinical studies there are probably much better targeted solutions than 23 andme Value of 23 and me is 1) brand and 2) data on large number of pts. 2) helps with finding genetic markers with low effect size which isn't really helpful for drug discovery and also finding rare variants with large effect, which is important, but there are better / cheaper ways of doing that (i.e. regeneron)