No one is incentivized to generate deep data on tumor subtypes, even lethal ones like TNBC that account for ~15% of breast cancer cases.
Proof: ask why no deep omics dataset exists for 1000 TNBC tissue samples, despite the obvious benefits.
It is astonishing what researchers have accomplished given the limited data and software they are armed with. Imagine how much faster they could move with good data.
To accelerate cure development, individuals can donate to Stanford instead of funding the 101st AI startup. Donations allow Stanford to generate more data, powering hundreds of cancer studies.
If you know senior OpenAI/Google execs who want to build cancer datasets, want to help with research [0], or wish to donate directly to Stanford [1], feel free to message me [2].
One goal is to help Stanford raise $250K - $5M for a deep omics dataset, starting with TNBC and NSCLC, which are cancers that mysteriously afflict young people. We receive no financial benefit; this is a passion project to patch a massive system bug that is more easily fixed by outsiders.
===
[0] Skills in CS, ML, math, bioinformatics, or virology? You can earn authorship on papers investigating under-studied questions in cancer research. Any of these is helpful -- no need for expertise in all.
[1] Donate directly here: give.stanford.edu/stanford-medicine. But please specify the department (Stanford Cancer Institute) and purpose, along with instructions to strictly fund data, ideally with 100% of proceeds and to request an audit trail of donation dollars. Overhead eats up the donation if you're not careful.
[2] twitter.com/panabee. Why do I follow biomedicine? There’s a surprising amount of ML overlap between computer vision and biomedicine.
===
twitter.com/panabee
Message for free Hotpot credits. HN is an invaluable source of knowledge. Happy to give back.