Interpreting single-cell and spatial omics data using deep neural network training dynamics
Jonathan Karin, Reshef Mintz, Barak Raveh, Mor Nitzan · 2024
Read Interpreting single-cell and spatial omics data using deep neural network training dynamics by Jonathan Karin, Reshef Mintz, Barak Raveh, Mor Nitzan (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-11659171.txt.
Description
Jonathan Karin, Reshef Mintz, Barak Raveh, and Mor Nitzan read the training process of a neural network to audit cell labels. Single-cell and spatial data sets sort cells into types, but those labels are judgment calls laid on noisy, sparse data. Some labels are simply wrong, and others blur real middle states.
Their framework, Annotatability, watches how hard a network finds each label while learning. Labels that fight training flag suspect groups or borderline cells. A companion method then builds embeddings that highlight the signal of interest. Across eight data sets, the approach caught bad labels, traced development and disease paths, and mapped cell diversity. The authors pitch it as a general lens for genomic data.
Catalog record
| CREATOR | Jonathan Karin, Reshef Mintz, Barak Raveh, Mor Nitzan |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 75 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |