Miguel Rocha

GRIDGENE identifies tissue compartments using transcript density in spatial transcriptomics data, enabling analysis without cell segmentation.

DeepMol automates the ML pipeline for molecular property prediction, achieving competitive results on benchmark datasets without manual feature engineering.

PENGUIN is a preprocessing tool for multiplexed spatial proteomics that normalizes signal-to-noise ratios and reduces noise without manual annotation or machine learning training.

The paper reconstructs and validates the first genome-scale metabolic model for Vitis vinifera, integrating RNA-Seq data to create tissue-specific models and explore ML-driven fluxomics analysis.

BioISO is a tool that aids in the manual curation of genome-scale metabolic models by reducing search spaces and improving debugging through recursive algorithms and flux balance analysis.

This study evaluates deep learning methods for predicting drug synergy in cancer by testing different architectures, data types, and feature selection strategies.

The paper presents a pipeline for reconstructing and evaluating context-specific human metabolic models using multi-omics data, validated on cancer cell lines and showing improved gene essentiality predictions.

merlin v4.0 improves genome-scale metabolic model reconstruction with new algorithms and tools, supporting multiple organisms and offering competitive performance.

This paper develops the first multi-tissue diel genome-scale metabolic model for Quercus suber, highlighting suberin biosynthesis pathways and enabling analysis of tissue interactions.

COSMOS integrates multi-omics data with prior knowledge to generate mechanistic hypotheses about biological processes in renal cell carcinoma.

HABIT is a webserver that uses machine learning to predict and analyze T cell epitopes for neoepitope discovery in immunotherapy.

The paper develops a consensus algorithm to improve tissue-specific metabolic models using multiple reconstruction methods and functional data, showing improved accuracy in phenotype predictions.