Cell Hashing Demultiplexer
Pool cells tagged with hashtag oligonucleotides (HTOs) into one 10x run, then watch a real thresholding classifier separate singlets, doublets and negatives in 3D HTO-count space — tune the doublet rate, ambient noise and threshold to see classification accuracy respond.
Cell hashing lets a lab pool several samples into one single-cell sequencing run by tagging each sample's cells with a distinct hashtag oligo (HTO) before mixing them into the same microfluidic lane. This simulator generates a realistic pool of cells under Poisson droplet-loading statistics, samples each cell's four-channel HTO UMI counts from signal-plus-ambient-noise distributions, and runs a real per-channel k·σ thresholding classifier to call every droplet a singlet (one sample), a doublet (two samples accidentally co-encapsulated), or a negative (no cell). Adjust the loading concentration, ambient noise level and classifier threshold to see the trade-off between catching true doublets and over-calling noisy singlets as false doublets, with live accuracy measured against the simulation's own ground truth.
Pool cells from four samples tagged with hashtag oligonucleotides (HTOs) into one sequencing run, then watch a real per-channel k-sigma thresholding classifier separate singlets, doublets and negatives in a live 3D HTO-count scatter, with Poisson droplet-loading statistics and tunable ambient noise driving classification accuracy.
3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install