Summary
Scientists often study how drugs work by looking at changes in gene activity, but measurements of small molecules called microRNAs, which also affect cells, are usually missing. The authors created a new method called MIRCID that guesses these important microRNAs from gene data, helping to better classify drug effects and find similar drug actions. Their method showed stronger and more consistent improvements than existing approaches that guess other regulators. This suggests that inferred microRNAs can be a useful way to understand drug effects when direct microRNA data is unavailable, although future work is needed to confirm these guesses with actual measurements.
What this means in practice
- •For pharmaceutical data teams: Enhance drug mechanism classification models by including inferred microRNA signals derived from gene expression data to improve similarity searches and predictions without direct microRNA measurements.
- •For biotech bioinformatics groups: Integrate inferred hub microRNAs into transcriptomic analyses to rescue weak drug effect signatures and gain better biological insights in drug perturbation experiments.
Authors
Xin Cao, Yigang Chen, Jiatong Xu, Ziyue Zhang, Xiang Cheng, Shenyu Wang, Yangyi Zhang, Xiaoxuan Cai, Shidong Cui, Zihao Zhu, Xiang Ji, Hsi-Yuan Huang, Yang-Chi-Dung Lin, Hsien-Da Huang
Abstract
Drug mechanism-of-action (MoA) modeling commonly relies on perturbational transcriptomes, but matched microRNA (miRNA) measurements are often unavailable. Inferred regulatory features offer a scalable way to reuse these data. Here, we present MIRCID, a framework comparing gene expression with inferred transcription factor (TF) activity and miRNA expression across pathway classification and similarity-based MoA retrieval. HubmiRNet infers 414 pan-cancer hub miRNAs (HubmiRs) from 977 L1000 landmark genes, achieving a Pearson correlation coefficient of 87.72\%; its 1,298-output variant also outperformed SiCmiR on the full-miRNA task (71.21\% versus 67.30\%). In the evaluated comparisons, miRNA augmentation provided more consistent gains than TF activity. Generic embedding controls showed model-dependent utility, while complementarity analyses identified a distinct, partially linearly recoverable representation that retained gene-derived structure. Illustrative rescue cases linked improved classification to biologically plausible miRNA patterns in samples with weak transcriptional signatures. These findings support inferred HubmiRs as a biologically informed recoding of transcriptomic data for perturbational drug modeling, while leaving recovery of measured perturbational miRNA responses to further validation.