AI Paper Atlas

Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning

Learning Biological Hierarchies in Single-Cell Foundation Models

Xiangyu Guo, Ricardo Henao

NeurIPS 2026 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Zero-shot robustness and transfer · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Introduces scOntoFM, a single-cell foundation model that embeds the hierarchical cell ontology into representation learning, using lowest common ancestor distances from the ontology.

Problem

Most single-cell foundation models ignore the taxonomic lineage relationships among cell types encoded in the cell ontology.

Approach

Embeds the hierarchical cell ontology into representation learning by using Lowest Common Ancestor (LCA) distances from the ontology to structure the learned cell representations.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add ontology-aware losses and offline sampling vs SCimilarity on zero-shot clusteringscOntoFM · zero-shot clustering, nine datasetsscIB: 0.7960 -> 0.7958No clear effect
ontology-aware pretraining vs baselines on structure preservationscOntoFM · zero-shot clusteringscGraph-LCA / scGraph-RWR: higher on most tissues (no numbers in text)Improved
scOntoFM at coarse annotation levelsscOntoFM · HIHA AIFI_L1scIB score: 0.760 (cellPLM) vs 0.681Hurt
scOntoFM at fine annotation levelsscOntoFM · HIHA AIFI_L2scIB score: 0.722 (cellPLM) -> 0.782Improved
fine-tune scOntoFM vs other scFMs on HIHAscOntoFM · HIHAaccuracy: 0.907 (cellPLM) -> 0.915Improved
novel cell-type identification with ontology-derived centroidsscOntoFM · 99 unseen cell typesaverage rank: 3.24 (scCello) -> 2.85Improved
gene function prediction with scOntoFM embeddingsscOntoFM · GO gene functionF1 micro: 0.913 -> 0.917Improved
scOntoFM vs Geneformer on APscOntoFM · GO gene functionAP micro: 0.922 -> 0.921No clear effect
spatial gene imputation with scOntoFM embeddingsscOntoFM · breast cancerPearson: 0.369 -> 0.388Improved
remove ontology alignment loss (ablation)scOntoFM · zero-shot clusteringscIB vs scGraph-LCA: scIB up, scGraph-LCA substantially down (no numbers in text)Mixed
remove dynamic margin ranking loss (ablation)scOntoFM · zero-shot clusteringscIB vs scGraph-LCA: scGraph-LCA up, scIB slightly down (no numbers in text)Mixed
offline ontology-guided sampling instead of random online samplingscOntoFM · 10% of pretraining datazero-positive anchors: 46.78% of anchors have zero positives onlineImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy