Prediction of Gene Regulatory Connections with Joint Single-Cell Foundation Models and Graph-Based Learning

What this source contributed

Records in this release that cite this source. A record can cite several sources.

• 2 pipelines
• 1 task
• 1 dataset
• 2 evaluations
• 2 results
• 5 evidence claims

Results

Values as the source printed them. Compare results only under compatible protocols, datasets and metrics.

Claims

Statements this source supports, each with where in the source it was checked.

• Reported AUROC for scRegNet (Geneformer backbone): Printed value 0.89 (Table 2, scRegNet (w/ Geneformer) row, hESC AUROC entry)
• Reported AUROC for scRegNet (scBERT backbone): Printed value 0.88 (Table 2, scRegNet (w/ scBERT) row, hESC AUROC entry)
• scRegNet (Geneformer backbone): description: Description Predicts gene-regulatory links by combining context-aware gene representations from the pretrained Geneformer single-cell foundation model with graph-based representations learned from known regulatory interactions. (Reviewed profile (2026-09-16): summary at Conclusions and Discussions (paragraph 1); Results/Performance on benchmark datasets (paragraph 2); architecture fact at Method/Graph-based learning with GNNs (paragraph 1); Method/Gene representations from foundation models (paragraph 1); description text also read at Introduction (scRegNet framework description); Abstract/Results)
• scRegNet (Geneformer backbone): uses model Geneformer: Uses model Geneformer (Methods scFM encoder and GNN encoder; Tables 2 and 3 scRegNet(w/ Geneformer))
• scRegNet (scBERT backbone): description: Description Predicts gene-regulatory links by combining context-aware gene representations from the pretrained scBERT single-cell foundation model with graph-based representations learned from known regulatory interactions. (Reviewed profile (2026-09-16): summary at Conclusions and Discussions (paragraph 1); Results/Performance on benchmark datasets (paragraph 2); architecture fact at Method/Graph-based learning with GNNs (paragraph 1); Method/Gene representations from foundation models (paragraph 1); description text also read at Introduction (scRegNet framework description); Abstract/Results)

Models, benchmarks and other records

Evidence

Source checking verifies the cited claim or transcription. It does not establish independent reproduction.