Matt Pelo

Billionaire John Morgan reveals only about 20 Americans have a REAL $1,000,000,000 liquid in the bank, and he's one of them… “I don't consider anybody a billionaire that doesn't have a real billion in the bank… I'm a billionaire. A real one.”

JUST IN: New York Democrats introduce bill requiring 10% discounts on all items at self-checkout

FDA is calling on patients, clinicians, and researchers to help identify drugs that could be repurposed to treat chronic and rare diseases and help address other unmet medical needs. Share your ideas to help advance new treatment options. fda.gov/news-events/press-an…

Not having kids is a life hack. I spend my summers in Europe, take cross-country road trips, and disappear off-grid on weekends. The money I’d spend on kids goes to investing, experiences, and freedom. Different priorities. Different definition of rich.

Everyone always asks how much a genome costs, but never how much one is worth. The Advances in Genome Biotechnology (@agbt) conference is celebrating its 25th anniversary in a few days. It's one of my favorite conferences - full stop. While I'm sad not to be able to attend this year, the FOMO has me reflecting a bit on next-gen sequencing (NGS). For 25 years, the industry has been drunkenly fixated on reducing the cost of sequencing a human genome. Nowadays, the amortized price of a leading NGS platform + the associated reagents needed for a 30x whole genome sequence (WGS) is <<$1,000, roughly six orders of magnitude cheaper than it was when AGBT got its start 25 years ago. Unfortunately, the value of a human genome sequence hasn't increased by the same fold in that time. For clarity, I'm talking about WGS for germline genetics, not research, not tumor sequencing, not NIPT, etc. This is a very focused point. The issue is that there just aren't that many high-risk monogenic variants that we can do something about immediately to alter medical management, which is where value comes from. Everyone knows about BRCA because it's one of the rare examples where, if you have a pathogenic/likely pathogenic (P/LP) variant, we can do something and be relatively certain we're helping you prevent an advanced breast cancer diagnosis. The bitter pill is that most diseases have vastly more complex genetic etiologies. This field has tried to push polygenic risk scores (PRS) for years, and they just haven't caught on clinically because they're not simple to interpret and are confounded by ancestral background. I'm also not aware of many controlled, HEOR analyses of system-wide implementation of PRSs for common conditions. I have been hoping that third-generation sequencing technologies (e.g., single-molecule) would start adding to the genome's value by tacking on historically recalcitrant genomic phenomena at scale - things like tandem repeats, phased variants, structural variants, and more. This might happen, but I'm skeptical it will add an order of magnitude more clinical value, which it needs to do to offset the still-higher price. The combination of better bioinformatics, swelling global sequence repositories, addition of orthogonal 'omics signals like proteomics, and more, are all helping add to the value of a human genome sequence. This is candidly what I've been curious about for the wave of groups pioneering DNA foundation models. Will the combination of longer-context windows (enough to grab distal regulatory elements) + other algorithmic innovation start to inject more momentum into the value conversation? If so, I haven't seen it yet. Thus, it's still a grand challenge, to link genetic variants and phenotypes, accurately and at scale, for everything from a single-nucleotide polymorphism to differential methylation status to large structural rearrangements. I won't say it's impossible, but it doesn't seem like we're making a massive dent yet. Very (very) happy to be proven wrong on this one.