Predicting dynamic expression patterns in budding yeast with a fungal DNA language model

Abstract

Predicting gene expression from DNA sequence remains challenging due to complex regulatory codes. We introduce a masked DNA language model pretrained on 165 fungal genomes related to budding yeast that learns conserved regulatory sequence motifs. Fine-tuning the LM on yeast RNA-seq data--including high-resolution transcriptional regulator induction time courses generated in this study--yielded Shorkie, a model that substantially improves gene expression prediction compared to baselines trained without self-supervision. Shorkie identified canonical transcription factor (TF) binding motifs and captured changes in their regulatory contributions across induction time courses. Furthermore, Shorkie accurately predicted variant effects, outperforming leading sequence-to-expression models in cis-eQTL classification and achieving high concordance with massively parallel reporter assays. Interpretability analyses revealed Shorkie's ability to resolve promoter dynamics, splicing signals, and temporal changes in regulatory motif usage. Together, these results support combining pretraining on related fungal genomes with RNA-seq fine-tuning to improve sequence-based prediction of gene expression in budding yeast and provide insights into noncoding variants.