Hesam Montazeri

iDriver: A patient-centric framework for genome-wide cancer driver discovery

iDriver improves cancer driver discovery by accounting for patient-specific mutation burden, identifying known and novel drivers across coding and noncoding regions and outperforming existing methods.

Reliable Evaluation and Learning in Multi-input Biological Association Prediction

The paper introduces a framework and model to evaluate and learn multi-input biological associations by eliminating shortcut biases, improving fairness and robustness in predictions.

Evaluating Single-Cell Perturbation Response Models Is Far from Straightforward

This study evaluates perturbation response models and finds existing metrics are unreliable due to scale and dimensionality issues, and complex models underperform simple baselines.

sCIN: A Contrastive Learning Framework for Single-Cell Multi-omics Data Integration

sCIN integrates single-cell multi-omics data using contrastive learning, outperforming existing methods on paired and unpaired datasets.

Uncovering Hidden Cancer Self-Dependencies through Analysis of shRNA-Level Dependency Scores

This paper introduces NBDep, a method that analyzes shRNA-level dependency scores to identify cancer self-dependencies, outperforming gene-level approaches in accuracy and statistical power.

Discovery of synthetic lethal interactions from large-scale pan-cancer perturbation screens

This study develops a statistical framework to identify synthetic lethal gene pairs from cancer perturbation screens and validates novel interactions in hepatocellular carcinoma.

APSiC: Analysis of Perturbation Screens for the Identification of Novel Cancer Genes

APSiC identifies novel cancer genes and non-genetic drivers from perturbation screens, revealing context-dependent oncogenic mechanisms and functional roles in specific cancers.