Dmitry Korkin

Large language model inference of macromolecular complex composition via model consensus and experimental data integration

The paper evaluates LLMs for retrieving gene components of macromolecular complexes and shows that integrating LLM consensus with experimental data improves accuracy.

Beyond gene length: Exon-intron architecture and isoform potential in the evolution of eukaryotic complexity

The paper shows that exon count increases beyond a certain point in eukaryotic evolution and saturates at ~10 exons per gene, suggesting it contributes to genomic complexity.

Real-Time Spatiotemporal Tracking of Infectious Outbreaks in Confined Environments with a Host-Pathogen Agent-Based System

This paper develops an agent-based model to simulate and predict infectious outbreaks in confined environments, showing it can accurately predict infection numbers and suggest improved containment protocols.

Robust and accurate diagnosis of infectious skin diseases from histopathology images by integrating deep learning and explainable AI

The paper develops SINDI, a framework combining deep learning and explainable AI to accurately diagnose infectious skin diseases from histopathology images, achieving high accuracy even on challenging grain-free images.

A spatial long-read approach at near-single-cell resolution reveals developmental regulation of splicing and polyadenylation sites in distinct cortical layers and cell types.

The paper introduces a spatial long-read sequencing method to study developmental splicing and polyadenylation changes in cortical layers and cell types, finding layer-specific and age-related regulatory patterns.

Biosurfer for systematic tracking of regulatory mechanisms leading to protein isoform diversity

Biosurfer identifies sources of protein isoform diversity by analyzing long-read RNA-seq data, revealing common mechanisms like alternative transcription start sites and 'ragged codons'.

Diverse intrinsic properties shape transcript stability and stabilization in Mycolicibacterium smegmatis

This study identifies transcript properties influencing stability and stabilization in Mycolicibacterium smegmatis under different growth conditions using mRNA degradation profiling and machine learning.

The Extent of Edgetic Perturbations in the Human Interactome Caused by Population-Specific Mutations

The paper analyzes population-specific mutations' impact on the human interactome using in-silico edgetic profiling of nsSNVs from the 1,000 Genomes Project.

Single-cell long-read mRNA isoform regulation is pervasive across mammalian brain regions, cell types, and development

The paper analyzes single-cell long-read mRNA isoform regulation across mouse brain regions, cell types, and development, finding pervasive isoform variability and cell-type specific splicing patterns.

Asymmetric trichotomous data partitioning enables development of predictive machine learning models using limited siRNA efficacy datasets

The paper develops a machine learning framework for predicting siRNA efficacy using a small dataset by applying trichotomous partitioning and random forest models, achieving better performance than linear models.

A Hybrid Deep Clustering Approach for Robust Cell Type Profiling Using Single-cell RNA-seq Data

This paper introduces a hybrid deep learning method for cell type clustering in scRNA-seq data, achieving performance comparable to supervised learning.

Structural genomics and interactomics of 2019 Wuhan novel coronavirus, 2019-nCoV, indicate evolutionary conserved functional regions of viral proteins

The paper uses computational methods to analyze structural genomics and interactomics of 2019-nCoV, comparing it to SARS coronavirus to identify conserved functional regions.

Boosting Gene Expression Clustering with System-Wide Biological Information: A Robust Autoencoder Approach

This paper introduces a deep learning-based clustering method for gene expression data that integrates biological information, outperforming existing methods on yeast and human datasets.

Enriching human interactome with functional mutations to detect high-impact network modules underlying complex diseases

The paper introduces DIMSUM, a computational framework that integrates GWAS, mutation effects, and PPI networks to detect disease-specific modules in the human interactome, showing improved biological relevance and disease association.

DISPOT: A simple knowledge-based protein domain interaction statistical potential

DISPOT estimates domain interaction propensities using SCOP family annotations from structurally resolved PPIs.

The genome of the soybean cyst nematode (Heterodera glycines) reveals complex patterns of duplications involved in the evolution of parasitism genes

The paper assembles and annotates the genome of the soybean cyst nematode, revealing duplications and horizontal gene transfer events in effector genes involved in parasitism.

Genome-wide prediction of bacterial effectors across six secretion system types using a feature-based supervised learning framework

The paper introduces PREFFECTOR, a feature-based supervised learning method to predict bacterial effectors across six secretion systems with high accuracy.

Multilayer view of pathogenic SNVs in human interactome through in-silico edgetic profiling

This study characterizes pathogenic SNVs and frameshift mutations by analyzing their edgetic effects on the human interactome, revealing differences in pleiotropy and interaction disruption.

Biological classification with RNA-Seq data: Can alternative splicing enhance machine learning classifier?

This study evaluates if isoform-level RNA-Seq data improves supervised machine learning performance in biological classification tasks compared to gene-level data.

Determining rewiring effects of alternatively spliced isoforms on protein-protein interactions using a computational approach

The paper develops a computational tool to predict the effects of alternative splicing on protein-protein interactions and demonstrates its utility in disease-related transcriptomic data.