Aritro Nath

This paper develops PhenoMap and PhenoSurv, models that infer tumor phenotypes from genomic data and improve cancer prognosis with interpretable deep learning.

The paper develops a combination therapy for ER+ breast cancer that uses an evolutionary double-bind strategy to suppress drug resistance.

The paper develops a prognostic model called ENDORSE to predict endocrine therapy response in advanced ER+ breast cancers using transcriptomic data and survival outcomes.

The paper shows that drug-sensitive breast cancer cells benefit from co-culture with resistant cells, which produce estradiol, leading to increased survival and resistance under treatment.

The paper shows that combining ribociclib and adavosertib reduces growth of mixed cancer cell cultures and limits resistance development.

The paper identifies lncRNAs as significant predictors of anticancer drug sensitivity, showing they can enhance drug response predictions beyond protein-coding genes.

The paper identifies cancer-specific eQTLs by modeling genotype-tumor purity interactions in heterogeneous tumor gene expression data, showing most risk variants affect tumor microenvironment cells.