Fatemeh Vafaee

This paper evaluates how much improvement in cross-cancer survival prediction is due to model selection methods versus actual concordance gains, finding evaluation effects can be as significant as reported gains.

This paper introduces optimal transport to explicitly reveal cell correspondences in single-cell multi-omics, improving integration and label transfer by identifying and removing ambiguous cells.

The paper identifies molecular features that distinguish functional from non-functional bacterial sRNA-mRNA interactions using integrative interactome and machine learning analysis.

The paper identifies cell-state-specific mechanisms of 5-azacytidine response in myelodysplastic neoplasms, showing distinct differentiation and transcriptional changes in cycling progenitors and quiescent stem cells.

This study uses single-cell multiomics to show that AZA response in MDS involves hematopoietic regeneration from mutated HSPCs and that AZA-refractive sub-populations contribute to disease progression.

The paper shows that pre-existing epithelial-mesenchymal (E/M) cell states determine differential phospho-signaling responses to EGF in NSCLC cells.

AlgoraeOS predicts drug synergy with dose-specific inhibition profiles and uncertainty estimates, achieving strong performance across diverse tissues and drug mechanisms.

This paper introduces a deep multiplexed imaging technique to profile up to 50 markers per circulating tumor cell, enhancing biomarker analysis for precision oncology.

Medlib provides a fast and feature-rich C/C++ library for exact alignment of Nanopore sequences using multiple edit distances, outperforming existing tools in speed and alignment completeness.

The paper predicts pregnancy outcomes in assisted reproduction by analyzing RNA from blastocoel fluid, showing higher RNA levels correlate with implantation failure.

SAFAARI integrates single-cell data, corrects batch effects, and annotates cell types using contrastive adversarial domain adaptation, outperforming existing methods in multiple benchmarks.

This study uses deep learning to compare human and mouse gene expression across 777K bulk RNA-seq samples, identifying genes with differing functional associations between species.

The paper develops GeneRAIN models trained on bulk RNA-seq data to predict gene attributes, showing they capture biological information beyond gene expression and outperform existing models.

The paper identifies cell-type-specific regulatory roles of a heptad of transcription factors in human hematopoietic stem and progenitor cells.

This paper introduces a method to convert single-cell RNA sequencing data into images for improved cell type classification using deep learning.

This study benchmarks eleven feature extraction methods for identifying blood-based transcriptomic biomarkers in cancer diagnosis, finding PLS-DA superior for classification and Ranger for robust feature selection.

This paper introduces scPSD, a pre-processing method that improves cell subtype separation in single-cell omics data by reducing complexity and enhancing cell-type discrimination.

This paper develops machine learning models to classify lentigo maligna and atypical intraepidermal melanocytic proliferation using reflectance confocal microscopy images.

This study identifies three lncRNAs consistently dysregulated across eight cancer types and shows they are involved in similar biological processes related to cancer progression.

This study benchmarks dimensionality reduction methods in scRNA-seq data using internal validation measures and DBSCAN, identifying LDA and PHATE as high-performing algorithms.

CANCERSIGN identifies and classifies mutational signatures in cancer genomes using raw mutation data, enabling clustering and analysis of ICGC datasets to discover novel signatures.