Sina Nassiri

TumorArchetypeR develops a modular framework to derive signature-based tumor subtypes, resolving fibrotic tumor heterogeneity and identifying clinically relevant subtypes with robust classifiers.

The paper compares scRNA-seq protocols for clinical trial samples, finding that GEM-X Flex with chop-fix outperforms standard protocols and snRNA-seq in fixed tissue and FFPE blocks.

This paper develops a distributed pipeline for immuno-oncology biomarker discovery using federated learning to integrate clinical and molecular data across institutions without sharing sensitive data.

This study identifies CXCR4-expressing tumor-associated macrophages in bladder cancer and shows that targeting them improves immunotherapy outcomes in mice and patients.

Mcadet improves feature selection in scRNA-seq by integrating MCA and community detection, showing better performance on fine-resolution and rare cell type datasets.

The paper introduces ICCELF, a framework for automated scRNA-seq cell type annotation that addresses class imbalance and data compositionality, improving annotation accuracy.

This study shows that pericytes restrict immune cell entry into the CNS by maintaining immune quiescence of brain endothelial cells, and their deficiency leads to increased leukocyte infiltration and severe neuroinflammation.

This study shows microglia control vessel calcification via TREM2, with ablation worsening calcification and revealing a new role in preventing ectopic calcification.