🚗GeneDrive.jl: Optimize Biological Vector Control,🌍aPhyloGeo: Phylogenetic Correlation Analysis,⛔StoPred: Protein Complex Stoichiometry Prediction🧪
Stay Updated with the Latest in Bioinformatics!
Issue: 109 | Date: 24 October 2025
👋 Welcome to the Bioinformer Weekly Roundup!
In this newsletter, we curate and bring you the most captivating stories, developments, and breakthroughs from the world of bioinformatics. Whether you are a seasoned researcher, a student, or simply curious about the intersection of biology and data science, we have got you covered. Subscribe now to stay ahead in the exciting realm of Bioinformatics!
🔬 Featured Research
The study investigates genomic signatures of historical natural selection potentially driven by ancient coronavirus epidemics. Using data from the China Kadoorie Biobank (CKB) and UK Biobank (UKB), the authors identified regions of long-range linkage disequilibrium (LRLD) enriched near genes encoding virus-interacting proteins (VIPs), specifically for coronaviruses in CKB but not UKB. These findings were validated using saltiLASSI, a selection-scan method robust to demographic confounders. No enrichment was observed for DNA virus VIPs or SARS-CoV-2 GWAS signals, supporting the hypothesis of ancient coronavirus-driven selection events localized to East Asia.
The study compares erythrocyte metabolism across lowlanders, acclimatized lowlanders, and native Tibetans using metabolomics and proteomics. Glycolytic flux remains stable, while the oxidative pentose phosphate pathway and glutathione levels are reduced under long-term hypoxia. Antioxidant enzyme profiles shift, with decreased superoxide dismutase and increased catalase. These changes suggest adaptive remodeling of redox and metabolic pathways in response to sustained high-altitude hypoxia.
PROCR encodes the endothelial protein C receptor, a key regulator of coagulation and vascular integrity. The study uses computational tools to identify six deleterious nsSNPs (N64T, F93L, R113C, P145L, R173H, R236W) affecting conserved residues and posttranslational modification sites. Structural modeling reveals disruptions in folding and protein interactions. Phenotype mapping links these variants to coagulation and cardiovascular disorders.
This study introduces ACU-Net, a modified U-Net architecture incorporating attention mechanisms for improved brain tumor segmentation on MRI slices. Using the BraTS 2018 dataset, the model was trained with dice and cross-entropy losses and evaluated using metrics like DSC and IoU. ACU-Net integrates attention gates in both encoder-decoder paths and skip connections to refine tumor boundary delineation. Comparative analysis with baseline models highlights its segmentation performance across tumor subregions.
Using data from over 550,000 Korean smokers, this study examined how oral health status, behaviors, and service utilization relate to smoking cessation intentions. Positive oral hygiene practices and unmet dental care needs were modestly associated with increased quit intentions. The findings suggest oral health may play a multidimensional role in motivating smoking cessation.
A pan-immunotherapy signature to predict intratumoral CD8+ T cell expansions | Nature Communications
Researchers developed a mouse model to track CD8+ T cell clonal dynamics across tumors and identified a transcriptomic signature in precursor exhausted cells predictive of clonal expansion. The signature correlated with treatment response in both mice and human PD-1 blockade therapy, and LAG-3 inhibition reactivated previously contracted clones.
This study profiled over 10 million cells from esophageal squamous cell carcinoma (ESCC) samples using mass cytometry. It identified immune deficits, such as reduced central memory T cells, and highlighted PD-L1+ macrophages and CD39+ T cells as correlates of immunotherapy response. Functional assays suggested potential for targeted immune modulation.
Researchers analyzed microarray datasets to identify 71 co-differentially expressed genes (co-DEGs) shared between psoriasis and T2DM. Functional enrichment and WGCNA revealed 33 candidate co-driver genes, with four (BEX5, EPHX2, GPRASP1, RBP4) validated in clinical samples. These genes were linked to immune pathways and regulatory networks involving TFs, miRNAs, and chemical modulators. Immune infiltration analysis and GSEA further characterized their roles in disease microenvironments.
Through bioinformatics analysis of transcriptomic datasets, this study identified 13 endothelial-mesenchymal transition-related genes common to endometriosis and recurrent miscarriage. Functional enrichment, immune infiltration profiling, and machine learning models highlighted shared molecular features and potential biomarkers.
This study presents "signal-strapping," a method using N-terminal histidine-appended signal peptides to identify metalloproteins via sequence searches. The approach led to the discovery of four bacterial protein families, including DUF4198 and DUF6702, with metal-binding capabilities. Structural and biochemical analyses confirmed metal coordination, particularly with nickel and copper ions. The method leverages conserved histidine motifs and genomic context to infer metal-related functions.
Fish isoallergens and variants: database compilation, in silico allergenicity prediction challenges, and epitope-based threshold optimization | Frontiers
This study compiled 79 fish isoallergen and variant entries from four allergen databases, focusing on ingestion-related reactions. Five in silico tools were evaluated for allergenicity prediction, with AllerCatPro 2.0 showing highest sensitivity. Epitope mapping and phylogenetic analysis of parvalbumins informed threshold optimization for allergen classification. The findings support improved prediction accuracy and highlight species- and isoform-specific allergenic variability.
🧰 Latest Tools
This study introduces ProtGram-DirectGCN, a two-stage framework that models protein sequences using globally inferred n-gram graphs and applies directed graph convolutional neural networks for protein-protein interaction prediction. Residue transition probabilities define graph edges, and embeddings are pooled via attention mechanisms. The method aims to offer a computational alternative to protein language models for link prediction tasks.
Source code is available here.
This study presents aPhyloGeo, a Python-based tool designed to correlate phylogenetic data with climatic variables. It enables visualization and analysis of genetic-climate relationships using user-defined datasets. The application supports integration of phylogenetic trees and environmental metadata to explore evolutionary patterns influenced by climate.
Source code is available here.
This study introduces PseudotimeDE-fast, a scalable method for detecting differentially expressed genes along cell pseudotime in single-cell RNA sequencing data. It addresses limitations in existing approaches by providing well-calibrated p-values and improved computational efficiency. The method was validated through simulations and real datasets.
Source code is available here.
AlphaDIA is a transfer learning framework developed for data-independent acquisition (DIA) proteomics, enabling feature-free analysis without reliance on spectral libraries. It leverages pre-trained models to enhance peptide identification and quantification across diverse datasets. The approach supports scalable proteomic profiling by reducing dependency on manually curated features.
Source code is available here.
MultiModalGraphics is an R package designed for annotated visualizations of high-dimensional biological data, including multi-omics datasets. It supports embedding statistical metrics like fold-changes and p-values directly into scatterplots and heatmaps. The tool integrates with Bioconductor packages such as MultiAssayExperiment and limma, enabling streamlined workflows from preprocessing to visualization. Case studies demonstrate its application across diverse multimodal datasets.
Source code is available here.
ASET is a modular pipeline for allele-specific expression analysis from RNA-Seq data, focusing on parent-of-origin effects. Built with Nextflow, it includes components for SNP-tolerant alignment, strand-specific read counting, contamination estimation, and gene annotation. The toolkit combines Julia scripts for statistical testing and R-based visualization, offering a reproducible and scalable solution for ASE studies.
Source code is available here.
VESNA is a Fiji macro for automated segmentation and skeletonization of 3D fluorescence images of vascular networks. It supports batch processing and requires minimal experimental preparation, making it adaptable across various tissue culturing methods. Demonstrations include organoids and hydrogel-based cultures, highlighting its utility in analyzing vascular structures in diverse experimental contexts.
Source code is available here.
GeneDrive.jl is a Julia-based software for optimizing genetic biocontrol strategies targeting mosquito vectors. It models temperature-sensitive dynamics and supports scenario-based testing aligned with local health and financial constraints. The tool enables planning and risk assessment for interventions like RIDL, demonstrated through case studies in dengue-endemic regions, and is designed for interdisciplinary use.
Source code is available here.
HCVTyper is a Nextflow-based pipeline for analyzing hepatitis C virus genomes from capture-based and metagenomic sequencing data. It includes modules for quality control, taxonomic classification, reference-based mapping, consensus generation, de novo assembly, and resistance annotation. The pipeline supports detection of mixed infections and low-frequency resistance-associated variants via integration with HCV-GLUE. It is containerized with Docker and provides detailed output reports.
Source code is available here.
Recommended by LinkedIn
StoPred is a deep learning method for predicting protein complex stoichiometry using protein language model embeddings and graph attention networks. It supports both homo- and hetero-oligomeric assemblies without relying on template structures or predefined compositions. The model captures subunit-level interactions and demonstrates improved accuracy over existing methods across curated and blind datasets. StoPred enables sequence- or structure-based inference of assembly composition.
Source code is available here.
🗞️ Community News
This blog talks about the launch of the world’s largest clinical trial aimed at finding treatments to slow or stop Parkinson’s disease. Led by UCL and Newcastle University, the £26 million trial will test multiple drugs in parallel, potentially saving years in drug development. It plans to recruit 1,600 participants across the UK and prioritizes drugs with existing evidence of promise. The trial is backed by major research and charitable organizations and hopes to set a new standard for future neurodegenerative disease trials.
This blog talks about how meditation can positively influence gut health through the gut-brain axis. It explains that the gut microbiome plays a key role in producing mood-regulating neurotransmitters like serotonin and GABA. Meditation, by reducing stress and promoting relaxation, helps balance this microbiome and improve overall mental and physical health. The article highlights the two-way relationship between brain activity and gut function, emphasizing the body's "second brain”.
This blog talks about a surprising discovery where a gene deficiency linked to obesity also offers protection against heart disease. Researchers found that individuals with MC4R gene mutations had lower cholesterol and better cardiovascular profiles despite being obese. The study challenges the traditional link between obesity and heart disease by highlighting the brain’s role in fat metabolism. These findings could pave the way for new therapeutic strategies targeting metabolic health.
At the ASHG 2025 meeting, researchers from the All of Us Program presented findings from Phase I of their long-read sequencing initiative, focusing on genetically diverse individuals. Using PacBio Sequel IIe, they identified nearly 1 million structural variants, including over 190,000 previously undetected ones. A phenome-wide association study linked 291 structural variants to 226 conditions, many absent in short-read data. Phase II will expand the study to additional populations, enhancing insights into genomic diversity and disease risk.
MIT researchers have discovered that the genome retains aspects of its 3D structure during cell division, challenging the long-held belief that mitosis erases chromatin organization. Using a high-resolution mapping technique called Region-Capture Micro-C (RC-MC), they identified persistent regulatory loops and newly described “microcompartments.” These structures strengthen as chromosomes compact, potentially preserving gene regulatory interactions across cell cycles. The findings offer new insights into how cells maintain transcriptional memory and regulate gene expression during division.
Researchers report a workflow that enables human genome sequencing in less than four hours from sample to variant calling. The approach integrates optimized sample preparation, rapid sequencing, and accelerated bioinformatics pipelines. It was tested on clinical samples to demonstrate feasibility for time-sensitive applications. The study outlines implications for emergency diagnostics and real-time genomic surveillance.
This article discusses research into cognitive symptoms experienced during menopause, commonly referred to as “brain fog.” The study examines changes in memory, attention, and executive function linked to hormonal fluctuations. Neuroimaging and cognitive testing were used to assess brain activity and performance. Findings contribute to understanding how menopause affects cognition and neurological health.
Scientists investigated how a ketogenic diet may influence brain aging and cognitive function. The study focused on metabolic changes in the brain, including ketone utilization and mitochondrial activity. Animal models were used to assess neuroprotective effects and age-related decline. Results suggest links between dietary fat intake and cellular mechanisms involved in brain maintenance.
The news is regarding a research which explored non-coding regions of the genome and their role in regulating gene expression linked to disease. Advanced sequencing and computational tools were used to identify regulatory elements and chromatin interactions. The study highlights how structural and epigenetic features contribute to disease susceptibility. It expands the understanding of genome architecture beyond protein-coding genes.
🗓️ Upcoming Events
This webinar introduces HiFi long-read sequencing for clinical laboratories, highlighting its ability to detect structural variants, repeat expansions, and complex genomic regions missed by traditional methods. It covers the integration of HiFi WGS and PureTarget into workflows for rare disease testing and carrier screening. A panel of experts discusses practical experiences in evaluating, validating, and implementing HiFi sequencing for diagnostic and translational applications.
This webinar explores integrated proteomics workflows that enhance throughput and data quality from limited sample input. It highlights advances in ion mobility separation, parallel accumulation-serial fragmentation modes, and Bruker’s TIMS-TOF system. The session covers sample preparation strategies and complementary data analysis tools aimed at improving proteome coverage and reproducibility.
🎓 Educational Corner
The article introduces Dirichlet regression as a principled approach for modeling plant community composition, addressing the limitations of traditional methods that ignore compositional constraints. It demonstrates how Gaussian process smooths within a Bayesian framework capture nonlinear ecological responses across environmental gradients while ensuring predictions respect the simplex constraint. Computational efficiency is achieved using approximate Hilbert space Gaussian processes, making the method practical for moderately large datasets.
Docker introduces an enhanced interactive prompt for docker model run, featuring readline-style editing, advanced keyboard shortcuts, and command history for improved usability. The update supports multi-line input, quick navigation, and efficient text editing, along with the ability to stop model responses instantly using Ctrl + c. These improvements aim to streamline local AI experimentation by providing a more intuitive and responsive CLI experience.
This post explains how Docker Model Runner integrates with Open WebUI to provide a full-featured local AI assistant experience. While Docker Model Runner handles model execution, Open WebUI adds capabilities like chat history, file uploads, voice input, and multi-model switching, all running locally for privacy and control. The extension simplifies setup by automatically provisioning containers and connecting to downloaded models, enabling a user-friendly interface for managing and interacting with local LLMs.
The guide details connecting OpenAI’s Codex to specialized MCP servers using Docker MCP Toolkit, enabling AI agents to interact with infrastructure tools like Neo4j. It covers one-click deployment of over 200 pre-built MCP servers, secure credential management, and consistent cross-platform configuration. A demonstration shows Codex leveraging MCP servers for tasks such as data modeling, graph creation, and query execution, illustrating how this integration transforms generic AI assistants into powerful, tool-aware systems for real-world workflows.
The article introduces the R-multiverse, a community-driven repository for R packages. It outlines the submission process, including GitHub/GitLab release creation and pull requests. Three packages—riem, geotargets, and weathercan—were submitted during a coworking session. The system supports automatic checks and quarterly production snapshots, allowing maintainers to update packages continuously.
This tutorial guides users through visualizing quasar redshift data using Polars, pandas, and Matplotlib within marimo notebooks. It covers spectral line analysis, Doppler shift concepts, and interactive dashboard creation. The course emphasizes data retrieval, cleaning, and dynamic visualization using reactive notebook components.
The article explains the distinction between Python’s .__repr__() and .__str__() methods. .__repr__() provides an official, unambiguous string representation for debugging, while .__str__() offers a user-friendly format. It demonstrates usage in built-in and custom classes, including formatting with repr(), str(), and f-strings.
Connect with Us
Stay connected and engage with us on social media for daily updates, discussions, and more!
· Website
· GitHub
📬 Subscribe
Don't miss an issue! Subscribe to the Bioinformer Weekly Roundup and receive the latest insights directly in your inbox.
We hope you enjoyed this week's edition of the Bioinformer Weekly Roundup. Feel free to share it with your colleagues and friends who share your passion for bioinformatics!
Disclaimer: The information provided in this newsletter is for educational and informational purposes only and does not constitute professional advice.
Contact: bioinformatics@zifornd.com
Copyright © 2025, Bioinformer Weekly Roundup. All rights reserved.