eganp

IIRC, this is still true if you exclude the spike protein and use a higher similarity threshold.

Why would one do that? Well, two reasons. First, the spike more variable as it is subject to greater evolutionary pressure as the major component of the viral surface. Second, in coronavirus reverse genetics typically the virus is segmented into multiple pieces, most of which compose the non-spike backbone and a subset which compose the spike protein. This is intentional to enable the swapping of spike proteins. So, if SARS-CoV-2 is spike-swap or spike variant assembly, a RaTG13-like virus could be the "backbone" of SARS-CoV-2.

Moreover, a hypothetical SARS-CoV-2 backbone could exist in the viral sequences list the WIV took down in Sept. 2019. They sequenced RaTG13 a few years prior. The lab's raison d'etre is collecting and sequencing coronaviruses poised for spillover. And, with modern DNA synthesis technology, it's not difficult to print out arbitrary backbone contigs for a viral assembly.

This doesn't test the "epigenetic information theory of aging."

First, there is no survival analysis. How is the mouse younger if it doesn't live longer? Similarly, the OSK "rejuvenated" mice display lower lean muscle mass.

Second, the causality is (willfully?) misinterpreted. The endonuclease used to causes DNA double-strand breaks does NOT directly alter the epigenome. Instead, it induces DNA-damage repair stress. One consequence, of many, is epigenetic (chromatin) dysregulation. DNA damage stress is well known to accelerate aging phenotypes. In fact, David published on how p53 stress from repeated DNA damage - using the same endonuclease setup - initiates a DNA damage response in turn promoting cell-cycle exit and cell elimination [0].

Third, cutting "non-coding" DNA in this case involves cutting specific ribosomes (cell translation machinery). Given that this pressure is constitutive, it's likely that these ribosomes evolve resistance to the nuclease by mutating functional sequences. However, the authors never assessed the mutation and function of these ribosomes.

Lastly, the in vivo AAV transduction efficiency isn't measured. This makes the OSK "rejuvenation" result hard to interpret. All cells get DNA damage (germline edit), but only transduced cells (<10% at best in whole organism) get some OSK exposure. Yet, the whole organism is "rejuvenated"? Is there a positive spill-over from OSK expression?

All the core claims about epigenetic information are either incorrect or grossly misleading. The perturbation, site-specific DNA damage, does not cause only loss of whole-cell epigenetic information. Hard to imagine how this got into Cell. I guess a big name and 20+ figures is all you need these days?

[0] https://doi.org/10.1016/j.devcel.2021.11.018

I'm surprised the insurance will pay for this. The judgment, and actions by Oberlin's administration, strongly suggest they were acting in bad faith. The college's behavior exceed negligence, the typical threshold in these sort of insurance contracts.

Pretty incredible Oberlin can get out of this considering what other insurers will do to avoid paying for incidents with cause.

$1M in seed funding, but at what valuation? I hope the terms are near market.

Never thought of this distinction before! Thanks for making this distinction.

This seems to hold in other contexts. People tend to tolerate price discrimination at an aggregate level (e.g. senior/student discounts) but loathe individual-level price discrimination (e.g. college tuition). I wonder if it's the unfairness or uncertainty

Paywall Bypass: https://archive.md/gi57Y

There a quite a few things missing from this paper that would make it a good a good drug discovery paper.

First, they didn't discover a drug - they found a hit. 30 days from target to hit using conventional high-throughput biochemical screening would take 2-4 months. So, this is 3x faster, but that's not the rate limiting step. Validation and in vivo studies will take >4 mo and 1-12mo respectively.

Second, if we take this as a "we found a hit" paper, I want to know how specific your hit is. This would be one of the major advantages of using AF2 - screen against related proteins with some structural or functional similarity. This is the time intensive and oft overlooked part of good in vitro screening campaigns. Potency is nice (although 9 μM isn't impressive), but ultimately selectivity is paramount when targeting a class of proteins with well conserved binding sites, like kinases. If they found a promiscuous CDK inhibitor that happens to hit CDK20, then I bet there are tons of previously reported promiscuous CDK inhibitors that will hit CDK20 too.

Third, this paper is surprising because it exploits none of the cool new things AF2 could enable. In addition to what you mention above, the authors could have tried to counter screen (much faster in silico!), find an allosteric inhibitor, identify a PPI/complex inhibitor, or take a leap by generating a SAR series in silico and validating a few selected compounds in vitro.

Overall, this paper seems both incremental and misdirected. Saving 2 months in the discovery phase, pre-IP, is worth ~0. Not sure anyone there has much experience developing drugs. Hits are nice, but rarely the hard part. However, a hit on a protein from a structurally divergent class would be a major accomplishment.

NIH likely funds 2/3rds of US biomedical research and ~1/3rd of global biomedical research by dollar value [0]. That link lists other big sources (ERC, MRC, US DoD, HHMI). My sense is that this landscape will change rather quickly with pandemic-inspired biomedical/defense initiatives and more private institutes in the US.

Moreover, NIH funding, due to its central role in supporting biomedical research, is often necessary for getting a tenured PI position. Many high caliber R1 schools require multiple R01 NIH grants for tenure. Often private funding sources are less valuable to universities because they pay lower indirect rates (<10% vs >40%) on sponsored research. As institutions are so dependent on NIH funding, the NIH can exert influence far beyond the research it funds directly.

[0] https://health-policy-systems.biomedcentral.com/articles/10....

Somewhat surprised asymmetric organocatalysis was awarded over asymmetric photocatalysis or photoredox chemistry. Photocatalysis is newer, has more demonstrated use cases, and holds more promise.

The list price will nearly always be a multiple of your cost with insurance. The issue here is that insurers have provisions in their network contracts that prevent providers from publically advertising lower prices than what's charged to the insurer, which mandates discounting. In many other industries, these most favored nation arrangements are considered anti-competitive and, thus, illegal.

Large hospital systems totally separate business and care, so it can be difficult to negotiate a discount. Smaller and independent providers are often more accommodating. However, you need to ask for a discount, you can't simply offer to pay cash (i.e. "If I pay cash, can I get 40% off"). The providers can counter your offer, but can't discount you off the bat or else they invite legal action from insurers.

Some insurers (looking at you United Health Care) have wholly owned PBMs. Thus, the PBM is an asset in its insurance trust. Any profit the PBM makes will appear as in increase in asset valuation for their insurance trust, or balance sheet in some cases.

However, insurance margins remain unaffected as the money did, in some sense, go out the door and is no longer in control by the insurer directly. It's not fraud. It's just anti-competivie self-dealing. So, no, it's not profits per se, but rather increased valuation via clever financial engineering.

Unfortunately this is pretty accurate. Pharmacy Benefit Managers (PBMs) are contracted by insurers to set the formulary, in effect the list of drugs the insurer will pay for. They negotiate to lower drug costs, usually via manufacturer rebates.

Where this is perverted, is that the PBMs have complex compensation schedules wherein they receive a percentage of drug rebates. Their goal is to secure the largest rebate possible. Thus, either negotiate more effectively OR pressure manufacturers to increase list prices and give larger rebates. In practice, the latter happens most often. Insurers don't mind high list prices because they get large rebates and it encourages people to stay insured. Net prices (pharma revenue) on off-patent insulin have declined consistently since mid-2000s [0].

However, insulin is inexcusably expensive due to legacy regulation about equivalence for biologics and sticky patient preferences. Even current net prices are 2-5x too high relative to what a competitive generic market could produce.

[0] https://www.fiercepharma.com/pharma/net-prices-for-insulins-...

I worked with a group trying to get a drug repurposing study running in June of last year. U01 submissions were often Sept-Aug with no cash until April 2021. To a degree, these NOTs and RFAs are deeply unserious and woefully underfunded. There are 10s of absurdly narrow sub-focuses (e.g. "Medical Consequences of Smoking and Vaping Drugs of Abuse in Individuals with HIV and COVID-19" really??) that are great add-ons to some existing research problem, but totally miss the point. It's not wrong to study these the ideas, but why prioritize over larger more impactful studies of drugs to treat ARDS or ALI?

Much like The Fed provided public and private debt support, these applications are more pork for existing investigators with deep knowledge of the NIH bureaucracy, but sadly likely of little relevance to the larger issue.

I'm glad we did. Vaccines were a home-run solution. In a crisis, what does academic R+D get you on the margin? Better testing for sure, but that wouldn't solve the crisis like vaccines would.

I will say the drug repurposing landscape was an absolute fiasco last year. Every funding opportunity wanted reams of data supporting use in COVID19 which takes time to generate. Yet, somehow on-patent anti-IL-6 drugs were tried again and again and again even in the absence of elevated IL-6.

UK and the NHS really got this right by centralizing repurposing efforts. In the US, there aren't more than ~50 experts in drug repurposing. Why didn't HHS/CDC round them up, lock them in a room, and have to rank-order molecules for proposed basket repurposing trials? We could have saved ~10-100k lives in the US with better drugs for early- and mid-stage disease, but nobody wants to fund a drug that can't generate an ROI. Truly a market, and institutional, failure here.

This study is looking at the appendix, not any brain tissue. Without considering symptoms, it's possible that altered PrP expression and/or folding in the appendix does not cause vCJD or result from exposure to PrP(Sc).

Also, these antibodies are not very selective for a given confirmation. Idk how this made it into the BMJ because standard in the PrP field is a protease resistance assay given the issues with Ab specificity.

Dangerous? The hypothesized connection is not remotely clear. In vitro transformation of human PrP into PrP(Sc) by bovine PrP(Sc) is a necessary for the causal model. The fact that a strong species barrier exists, in vitro and in vivo, points towards a spurious association.

It's understandable why one might infer from the correlation that there's a relationship but this is confounded by observation. The baseline expectation for CJD in a population is 1-2 cases/million [0]. vCJD cases in the UK never exceeded 30/yr [1, Fig 2. a]. Yes, vCJD occurs primarily in younger people - but that's circular as vCJD is defined as CJD-like disease with early onset. If cross-species transmission were common, one would expect a similar spike in age >40.

The incidence of vCJD, even at its peak, was exceedingly rare. Many things, such as observation bias, pollution, or rare side-effects from faddish street drugs could be at fault here. It's tempting to jump to conclusions, a la "eating fats makes you fat," but it's totally unwarranted in the face of the in vitro and in vivo data demonstrating strong species barriers.

[0] https://www.cdc.gov/prions/cjd/occurrence-transmission.html [1] https://link.springer.com/article/10.1007/s00401-020-02153-7

Intracranial inoculation of infected tissue is the old-school way to assay for prion activity. Sometimes it takes a few years to get an answer though, especially in cross-species assays [0].

[0] https://link.springer.com/chapter/10.1007/978-1-4614-5338-3_...

This probably isn't possible as the folding of PrP is not encoded in the mRNA of PRNP. The fold of PrP is a post-translational phenomenon, much like its glycosylation. This is true for many mammalian proteins, especially cell-surface proteins.

The species barrier in prion transmission appear very strong [0]. Transmitting a prion from a hamster into a mouse often takes multiple passages of direct in brain inoculation. In these closely related species, perhaps it's less strong. It's doubtful that CWD can transmit to humans at typical exposure levels. In vitro, conversion of human PrP is possible, but exceedingly difficult.

[0] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2736662/ [1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5667731/

The risk of cow->human BSE transmission is infinitesimal. Species barriers in prions are surprisingly strong and, for practical purposes, nearly impermeable [0]. In fact, the Prion Protein (PrP) in humans causes multiple distinct diseases including CJD, Kuru, and FFI. Infectious prion confirmations are probably highly influenced by glycosylation [1,2]. Even if this testing wasn't god awfully slow, it's unclear one learns anything from it.

[0] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2736662/ [1] https://www.cureffi.org/2013/05/05/prion-protein-n-linked-gl... [2] https://www.biorxiv.org/content/10.1101/2021.02.14.431014v2

Clearly open access has costs, but how does Nature Photonics get to $40k/paper?? Is that how much revenue they expect to derive from the lifetime of the paper?

It's a tough lift to demonstrate that all such modified-mRNA vaccines are safe and likely effective. This trial is a good validation of off-target safety, and that's a huge achievement. However, you still have on-target safety concerns (e.g. cross reactivity). Then, for each pathogen, the selected antigen expressed needs to 1. generate an antibody response and 2. provide protection from infection.

Antibody titers can de-risk 1 pretty effectively. But, without human challenge trials, you're stuck waiting for nature to test 2. Even using non-human primates as the model to test infection has many pitfalls as the species barrier is often strong with infectious diseases. Human immunology is often idiosyncratic when compared with closely related primates.

Hey refoundable - great interview! Love Jack's characterization of Alzheimers as chemical roulette.

Just wanted to let you know there's a typo in your first figure. "Thalyidamide" should be "thalidomide" [0]. It's also been sold as Thalomid.

[0] https://en.wikipedia.org/wiki/Thalidomide

This is a nice essay but I think it misses 3 key points:

1. FTC returns are predictable. The returns to taking an unbeaten path is anything but. Further, how will your peers and the world view you as CEO of New Idea Inc vs. L4 at Google? 2. Prestige is a cheap correlate, and thus substitute, for aptitude. Outside of Tech, verifying aptitude and fit is like trying to pay with physical gold whereas prestige is like paying with Amex Platinum. 3. Consulting and Finance offer exposure to a swath of otherwise opaque industries and opportunities. This appears to be especially true in industries driven by culture fit and trust. Why do many e.g. biotech companies eschew an MD/PhD CEO for an ex-MBB 'professional CEO'? Because the CEO exists raise money and that requires immense trust and penetrating opaque networks.

Perhaps returns to prestige are overrated, but I'll take the prestige 10/10. In my experience, the previously closed doors suddenly open themselves for you.

How much of this compensation is equity appreciation? Is it realistic to expect this 10 years from now?

Not like interest rates can get lower or search could get less competitive.

Ok let's suppose that consulting salaires hit $1M by age 40. What fraction of new hires make senior consultant? Partner?

If your firm is operating on leverage of 20:1, sure you can make $2-20M annually. But, realistically, what are your odds on the margin of making partner 25 years down the road?

I wish I knew this coming out of school. I think this is also another reason why people prefer prestigious schools. Those who own and operate these cash machines tend to send their children to elite schools.

I can't even begin to count the number of people I've met at Stanford whose parents own "X," where X is a little publicized, privately held business with >$10M in revenues.