Extended Data Fig. 5: Ablation and enrichment analysis of GET-MPRA.
From: A foundation model of transcription across human cell types
a. Scatter plot of lentiMPRA readout versus GET-MPRA prediction (top), observed ATAC signal (middle) and sum of GET-MPRA prediction and observed ATAC signal (bottom). b. Promoter (top) or ATAC peak (bottom) elements are gated into four sub-categories, respectively, based on high (+) or low (−) in Prediction (cutoff=1) or Observation (cutoff=0.5). c. Histone mark enrichment analysis of promoter (top) and peak (bottom) elements respectively using ENCODE K562 ChIP-seq data. d. Transcription factor binding site enrichment analysis of promoter (Top) and peak (Bottom) elements respectively using ENCODE K562 ChIP-seq data. Fisher exact test was performed. Tests with a p-value < 0.05 are shown. Color shows log10 (Fold enrichment). For transcription factors, the variance of fold enrichment across four groups was calculated, and the top 50 TFs are visualized.
