schubert – Institute of Bioinformatics

SCHUBERT RESEARCH GROUP

We study cancer gene expression using advanced statistical methods to uncover new biological insights.

Research

Transcriptomic Footprints of Signaling Pathway Activation

We infer pathway activity from downstream perturbation-responsive gene signatures rather than from the expression of pathway members themselves, which helps overcome a key limitation of transcriptomics-only pathway mapping: that signaling is often controlled by post-translational regulation. By aggregating a large compendium of perturbation experiments into consensus Pathway RespOnsive GENes (PROGENy), we can better explain drug sensitivity and complement/refine mutation-based biomarkers.

M Schubert et al., Nat. Commun. 9, 20 (2018). doi.org/10.1038/s41467-017-02391-6
F Iorio et al., Cell 166, 740–754 (2016). doi.org/10.1016/j.cell.2016.06.017

Statistical modeling of gene expression

A central theme of our work is the use of advanced statistical models to separate expected gene-expression behavior from biological covariates. For example, we modeled RNA-seq counts with Bayesian negative binomial regression to decompose copy-number–driven scaling from residual deregulation. Applying this framework, we identified a specific vulnerability of CCND1-RBM14 co-amplifications to radiation treatment in colorectal cancer patients.

V Rendo*, M Schubert* et al., Nat. Commun. 16, 1077 (2025). doi.org/10.1038/s41467-025-56301-2

Inflammation and Chromosomal Instability

We study inflammatory processes in cancer, particularly those related to Chromosomal Instability (CIN). This goes from close collaborations with experimental scientists that study a particular CIN system (e.g. Mps1 inhibition or GEMMs) to large-scale single-cell human cancer atlases where we improve the accuracy of copy number inference for rare events.

C Hong, M Schubert et al. Nature 607, 366–373 (2022). doi.org/10.1038/s41586-022-04847-2

* Co-first author, ^# Corresponding author. For a full list of publications, please visit the Google Scholar page.