Fayyaz Minhas

Discovering Prognostic Risk Strata and Heterogeneous Progression Routes from Cross-Sectional Survival Data

GOTSurv models patient-specific disease progression from cross-sectional survival data, identifying heterogeneous risk strata and individualized progression trajectories without predefined risk cut-offs.

DeepPathway: Predicting Pathway Expression from Histopathology Images

DeepPathway predicts pathway expression from H&E images using contrastive learning, validated on prostate cancer and brain tumor datasets.

Whole slide image analysis of the endometrial decidual reaction reveals multiscale perturbations associated with miscarriage

The paper develops a deep learning tool to analyze endometrial histology and identifies morphological correlates of miscarriage risk.

DCAFA: Differential Community Abundance and Feature Analysis for Histological Images

DCAFA analyzes histological images to identify compositional shifts and feature associations in biomedical data using a regression-based framework.

GOTFlow: Learning Directed Population Transitions from Cross-Sectional Biomedical Data with Optimal Transport

GOTFlow infers directed population transitions from cross-sectional biomedical data using optimal transport, revealing biologically meaningful dynamics in endometrial remodelling, breast cancer risk, and prion disease.

Buyer Beware: confounding factors and biases abound when predicting omics-based biomarkers from histological images

This paper shows that deep learning models for predicting biomarkers from histology images are confounded by interdependencies among biomarkers, leading to less accurate and less reliable predictions.

Star-Motifs: Revealing Single-Cell Spatiotypes from Routine Histology Using Star-Convex Neighborhoods

Star-Motifs identifies spatial patterns in histology slides to reveal clinically relevant tumor microenvironment features and survival correlations.

Predicting Degradation Potential of Protein Targeting Chimeras

The paper presents a machine-learning pipeline to predict PROTAC degradation potential using compound fingerprints and a graph neural network, achieving better performance than existing methods.

Predicting small-molecule inhibition of protein complexes

The paper introduces a graph neural network to predict small-molecule inhibition of protein complexes using compound and protein structure inputs, achieving high accuracy on benchmark datasets.

Quadratic Programming Data Descriptors for Abnormal Beat Detection in ECG Recordings

This paper introduces a novel one-class classifier for abnormal beat detection in ECG using wavelet domain features and evaluates it on the MIT-BIH arrhythmia database.