Bioinfospace Blog
-
Rahul Satija's Eight Major Single-Cell Breakthroughs of 2025
Rahul Satija's 8 single-cell breakthroughs of 2025 from the 9th Single Cell Genomics Day — covering both experimental and computational advances from the Seurat team.
-
You Really Do Need to Know a Little Nextflow!
Why Nextflow beats custom shell scripts for bioinformatics: resumable pipelines, automatic parallelism, and reproducible workflows — with a practical single-cell example.
-
Linux Tips: How to Run Commands in the Background
Keep long bioinformatics jobs running after logout using nohup, screen, and tmux — so server disconnections and closed terminals never lose hours of compute.
-
2025 Nobel Prize in Physiology or Medicine: How Was the Immune System's 'Safety Guardian' Discovered?
The 2025 Nobel Prize in Physiology or Medicine awarded to Brunkow, Ramsdell, and Sakaguchi for discovering the FOXP3/Treg axis — how regulatory T cells prevent autoimmune disease.
-
Weekly Paper Feed: February 12–18, 2026
This week's top bioinformatics preprints: genome-wide Perturb-seq at scale, foundation model benchmarking, a basal ganglia epigenomic atlas, 3D spatial deconvolution, and ML-guided spatial omics.
-
Coding Journal Club: Running ViVAE on Tabula Sapiens Bone Marrow
We ran the official ViVAE tutorial on 12,297 bone marrow cells. It beat t-SNE and UMAP on global structure (0.569 vs 0.419/0.410) but came last on local structure. Working code, real numbers, and the install fixes you will need.
-
Weekly Paper Feed: February 11, 2026
This week: ovrlpy tackles 3D spatial doublets, Seq-Scope-X breaks optical resolution, Nucleotide GPT reveals what genomic language models really learn, SAFAARI for single-cell integration, and AGED for reference-free deconvolution.
-
Hallmarks of Cancer 2026: The Systemic Era
The updated Hallmarks of Cancer (2026) framework adds Systemic Interactions and Hallmark-Conveying Cells to the core principles. Here's what bioinformaticians need to know.
-
Journal Club: Multi-Omics Integration Across the Cancer Continuum – From Early Detection to Tumor Evolution
This month, we dissect three Nature Communications papers showcasing how integrated multi-omics—from plasma fragmentomics to single-cell atlases—is transforming cancer diagnosis, subtyping, and our understanding of tumor plasticity.
-
Journal Club: From Genomes to Phenotypes—How Computational Integration and Precision Models Are Reshaping Biomedical Discovery
This month, we explore three papers that, at first glance, span cancer genomics, bioinformatics methods, and antibiotic resistance. Yet, a deeper look reveals a unifying narrative: the critical need for causal, mechanism-driven models in biology, powered by integrated computational frameworks, to tackle complex problems from tumorigenesis to antimicrobial resistance.
-
Journal Club: Single-Cell Transcriptomics Decodes Fate Decisions in Regeneration and Disease
How do cells decide between healing and scarring? This month, we explore three papers using single-cell RNA-seq to uncover dynamic gene switches—from SOX9 in kidney fibrosis to brassinosteroids in plant roots—and the computational tools powering these discoveries.
-
Beyond the Code: How BioDeepFuse Uses AI to Decode the Genome's Hidden Managers
A new hybrid AI tool, BioDeepFuse, is cracking the code of non-coding RNA, offering a powerful and accurate way to classify these crucial genomic regulators.
-
Teaching AI Genomics with a DNA Mirror: How Self-GenomeNet Cracks the Data-Scarcity Problem
What if you could train powerful genomic AI models with 90% less labeled data? A new self-supervised method, Self-GenomeNet, uses DNA's own symmetry to learn the language of life.
-
Deep-N6mA: A New Deep Learning Tool Shines a Light on the Epigenetic Dark Matter
A new deep learning model, Deep-N6mA, is setting a new standard for accurately detecting a crucial but elusive DNA modification, opening doors for epigenetic discovery.
-
Deep Learning Decodes Biology: How Neural Networks Are Revolutionizing Genomics and Imaging
The data deluge in biology has outgrown traditional tools. Discover how deep learning is uncovering hidden patterns in our genomes and cells.
-
Beyond the Genome: How Deep Learning is Weaving the Tapestry of Human Health
A flood of molecular data holds the key to personalized medicine. We explore how deep learning is finally stitching it all together.
-
Beyond the Swab: How Multi-Omics Uncovered COVID-19's Molecular Fingerprint
A new study cuts through diagnostic uncertainty, identifying two key proteins that could rapidly diagnose SARS-CoV-2 and reveal hidden asymptomatic cases.
-
Beyond Viral Suppression: Multi-Omics Reveals Hidden Metabolic Risk in Treated HIV
New research uses network analysis to uncover distinct metabolic profiles in people with treated HIV, revealing a hidden at-risk group that could benefit from personalized interventions.
-
How to: use data.table for more efficient data management
data.table accelerates R by 10-100x through in-place operations and minimal memory overhead. The DT[i, j, by] syntax elegantly handles large genomics datasets that would cripple base R or dplyr.
-
Cell Annotation: AI-based
Foundation models like scGPT automate cell-type annotation by learning directly from single-cell RNA data. This approach annotates thousands of clusters in seconds while eliminating manual marker lookup.
-
Microarray Data Analysis
Load CEL files with affy, normalize using RMA, and identify differentially expressed genes with limma—your complete Bioconductor workflow for mining GEO's millions of microarray datasets.
-
Set up a remote server
Move your bioinformatics work to a remote server: configure SSH access, secure with firewalls and keys, containerize with Conda and Docker, and optimize storage with RAID for long-running analyses.
-
Alternative to cellranger: STARsolo
STARsolo replaces Cell Ranger for 10x scRNA-seq quantification—10× faster, open source, cross-platform, with identical results and complete parameter flexibility. (161 characters)
-
GSEA, GSVA and ssGSEA
GSEA compares two conditions, GSVA transforms expression to pathway scores across samples, ssGSEA scores pathways per individual sample. Choose based on experimental design and biological question.
-
Running Cell Ranger
Step-by-step guide to running Cell Ranger for 10x scRNA-seq analysis. Covers mkfastq, count, and multi pipelines with best practices for reference genome setup.
-
BD Rhapsody and Seven Bridges
BD Rhapsody's microwell-magnetic bead capture excels at sequencing fragile cells better than droplet methods. Analyze results via Seven Bridges (paid cloud) or Rhapsody Pipeline (free, local).
-
Cell Annotation: Canonical markers
Use canonical markers—experimentally validated genes with high cell-type specificity—to annotate single-cell clusters. Includes curated lists for immune, epithelial, and stromal cell types with R validation code.
-
Cell Annotation: Reference Mapping
**Automatically label cell types by mapping to reference atlases.** SingleR, Azimuth, and Symphony automate what would take hours of manual marker inspection—choosing the right tool depends on your reference quality and…
-
SCENIC tutorial
SCENIC infers transcription factor activity from scRNA-seq to identify master regulators. Three-step workflow: GRN (co-expression), CTX (motif enrichment), AUCell (regulon scoring).
-
Get Shiny Server on DigitalOcean
Deploy R Shiny apps on a Linux server using open-source Shiny Server for cost-effective, private hosting with custom domains. Learn installation, multi-app configuration, and per-user directory setup—all for $5–10/month.
-
How to: Use Parallel Loop in Bash
Use bash `&` or GNU `parallel` to process FASTQ files in parallel instead of sequentially, cutting analysis time dramatically. Remember: N parallel jobs × memory-per-job must not exceed your available RAM.
-
Deploy oPOSSUM3
oPOSSUM3 analyzes transcription factor binding sites (TFBS) in gene lists using legacy Perl/CGI infrastructure. Docker-based deployment and manual installation strategies for modern bioinformatics.
-
TCR/BCR Reportoire Analysis
V(D)J analysis with MiXCR or Cell Ranger reveals immune cell diversity. Visualize clonality and TCR/BCR repertoire dynamics in R using immunarch for adaptive immune research.
-
Velocyto/RNA velocity
RNA velocity predicts cell state transitions by comparing nascent vs mature mRNA. Use Velocyto for intron counting and scVelo for trajectory inference in dynamic biological processes.
-
Download public sequence data from SRA
Use sra-toolkit (prefetch and fasterq-dump) to download raw sequencing data from the Sequence Read Archive. For 10x single-cell data, split files correctly or download original BAMs.