Mohammad R. K. Mofrad

This study investigates the identifiability of metabolic resilience from sparse longitudinal metabolomics data using simulations and stool metabolomics from the Integrative Human Microbiome Project.

The paper evaluates how well microbiome measurements from different intestinal regions predict intervention decisions compared to stool samples.

The paper identifies force gates in the LINC complex using molecular dynamics simulations, showing how specific molecular contacts regulate mechanical force transmission across the nuclear envelope.

Measurement-induced decision loss (MIDL), host benefit forgone versus perfect observation, grows linearly with microbiome information loss in a binary model, then plateaus once the optimal policy becomes signal-independent.

This study shows lipid nanoparticles form protein coronas through lipoprotein fusion, not surface adsorption, using cryo-EM and simulations.

This paper reports on the fourth CAFA challenge, showing improvements in Gene Ontology term prediction but highlighting challenges in new evaluation settings.

The paper identifies an expanded HP1a-binding motif, HACs, and shows their role in heterochromatin assembly through molecular interactions and simulations.

KODA uses LLMs and knowledge graphs to identify antimicrobial drug targets in the gut microbiome by analyzing metabolic pathways and microbial interactions.

SIMBA-GNN predicts microbial abundances and cross-feeding relationships by integrating metabolic simulations with a graph neural network, achieving high performance.

MetaBiome integrates agent-based modeling and metabolic networks to simulate spatial regulation in gut microbial communities, revealing how microbes adapt to environmental conditions.

A 40B-parameter DNA model trained on 9.3 trillion tokens scores coding and noncoding variants zero-shot, exposes interpretable sequence features, and writes chromosome-scale sequences.

This paper demonstrates that adding small molecules to plasma enhances nanoparticle protein corona diversity, significantly increasing the number of detectable proteins in plasma proteome profiling.

The paper introduces cdsFM, a suite of codon-resolution large language models that predict the functional consequences of synonymous codon choices and identify pathogenic variants.

This paper fine-tunes protein embeddings to improve interpretable function prediction, showing that fine-tuned models outperform pre-trained ones in both known and novel function classification.

GO Bench provides a benchmarking hub for evaluating machine learning models in protein function annotation using Gene Ontology terms.

The paper develops function-conditioned generative models to predict functionally important protein residues using Gene Ontology data and protein language models.

TripletProt improves protein functional annotation using Siamese networks, achieving better F1 scores than existing methods in sub-cellular localization and gene ontology prediction.

EpitopeVec improves linear epitope prediction using deep protein sequence embeddings, achieving better accuracy and AUC than existing methods.

The paper reports results from the CAFA3 challenge, showing improved computational protein function prediction and new annotations for over 1000 genes through experimental screens in C. albicans, P. aureginosa, and Drosophila.

The paper determines the structures of engineered knottins 2.5F and 2.5D in complex with integrin αVβ3 and reveals structural mechanisms underlying their differential binding to αVβ3 and α5β1.

This study develops a deep learning model for protein secondary structure prediction using primary sequences, achieving high accuracy on a challenging dataset.

This paper introduces PPE segmentation for protein sequences, enabling discriminative motif discovery (DiMotif) and sequence embedding (ProtVecX) with improved performance in motif detection and protein classification.

This paper introduces DiTaxa, a method for analyzing 16S rRNA sequences using nucleotide-pair encoding to improve microbiome phenotype classification and biomarker detection.

MicroPheno predicts environments and host phenotypes from 16S rRNA gene sequencing using k-mer features from shallow sub-samples, outperforming OTU-based methods in accuracy and efficiency.