Biology subjects

Shahid, S. A.

Publications and source records attributed to Shahid, S. A..

3 recordsLinked to original sources

bioRxiv2025

An integrative machine learning approach identifies the centrality of ferroptosis, cuproptosis, and immune pathway crosstalk for breast cancer stratification and therapy guidance.

Breast cancer (BRCA) is a leading cause of cancer-related mortality in women, characterized by marked heterogeneity in molecular subtypes, immune microenvironment, and therapeutic response. Current gene expression classifiers often lack mechanistic grounding, limiting their clinical utility. Using an integrated machine learning approach, we identified a four-gene panel, FOXO4, EGFR, FGF2, and CDKN2A, capturing convergent dysregulation across ferroptosis, cuproptosis, and immune pathways. The panel reflects not only redox and proliferative dysregulation but also distinct immune microenvironmental patterns, with FGF2 linked to stromal remodeling and CDKN2A correlated with adaptive immune responses, underscoring its biological integration into BRCA pathology. This panel is rooted in recurrent dysregulation of oxidative stress control (FOXO4), proliferative and angiogenic signalling (EGFR, FGF2), and cell-cycle-immune interfaces (CDKN2A), linking classification to central regulatory mechanisms. The model achieved 97-98% test accuracy (AUC 0.99) for tumour-healthy discrimination. Our findings reveal transcriptional convergence between redox-metabolic and immune-escape programs in BRCA. We propose a compact and interpretable panel with translational potential, offering a minimal, mechanism-informed diagnostic framework for clinical deployment in breast cancer for precision oncology.

cancer biology↗

bioRxiv2025

Minimal Gene Signatures Enable High-Accuracy Prediction of Antibiotic Resistance in Pseudomonas aeruginosa

Antimicrobial resistance (AMR) in Pseudomonas aeruginosa poses a critical global health challenge, with current diagnostics relying on slow, culture-based methods. Here, we present a ML framework leveraging transcriptomic data to predict antibiotic resistance with high accuracy. We applied a genetic algorithm to 414 clinical isolates to identify minimal, highly predictive gene sets ([~]35-40 genes) distinguishing resistant from susceptible strains for meropenem, ciprofloxacin, tobramycin, and ceftazidime. Automated ML classifiers trained on these sets achieved accuracies of 96-99% on test data (F1 scores: 0.93-0.99), surpassing clinical deployment thresholds. Multiple distinct, non-overlapping gene subsets exhibited comparable performance, indicating that resistance acquisition broadly impacts the expression of diverse regulatory and metabolic genes. Comparison with known resistance markers from CARD and operon annotations revealed a substantial number of previously unannotated clusters, highlighting significant knowledge gaps in current AMR understanding. Mapping these genes onto independently modulated gene sets (iModulons) revealed transcriptional adaptations across diverse genetic regions. Overall, this study presents a streamlined machine-learning workflow for transcriptomic data and offers a pathway toward rapid diagnostics and personalized treatment strategies against AMR.

systems biology↗

bioRxiv2024

An Automated Machine Learning Framework for Antimicrobial Resistance Prediction Through Transcriptomics

The emergence of antimicrobial resistance (AMR) poses a global threat of growing concern to the healthcare system. To mitigate the spread of resistant pathogens, physicians must identify the susceptibility profile of every patients infection in order to prescribe the appropriate antibiotic. Furthermore, disease control centers need to be able to accurately track the patterns of resistance and susceptibility of pathogens to different antibiotics. To achieve this, high-throughput methods are required to accurately predict the resistance profile of a pathogenic microbe in an automated manner. In this work, a transcriptomics-based approach utilizing a machine learning framework is used to achieve this goal. The study highlights the potential of using gene expression as an indicator of resistance to different antibiotics. Results indicate the importance of starting with a high-quality training dataset containing high genetic diversity and a sufficient number of resistant samples. Furthermore, the performed analysis reveals the importance of developing new methods of feature reduction specific to transcriptomic data. Most importantly, this study serves as a proof-of-concept to the potential impact of deploying such models to reduce the mortality rate associated with AMR.

systems biology↗