Bioinformatics Pipelines – Jubayer Hossain

End-to-end bulk RNA-seq Quantification Pipeline using Salmon

RNA-Seq

A complete pipeline for pseudo-alignment and quantification of bulk RNA-seq data using Salmon — from raw sequencing reads to transcript-level abundance estimates ready for downstream differential expression analysis.

Input: Raw FASTQ files, reference transcriptome

Output: Transcript/gene-level count matrix, quantification summary

Python Salmon tximeta DESeq2

GitHub

nf-core/rnaseqmeta: A Nextflow Pipeline for Meta-Analysis of RNA-seq Datasets

RNA-Seq

A Nextflow-based pipeline for reproducible meta-analysis of multiple RNA-seq datasets — automating sample retrieval, quality control, quantification, batch correction, and integrated differential expression analysis across studies.

Input: Multiple RNA-seq datasets (FASTQ or SRA accessions), sample sheets

Output: Integrated count matrix, cross-study DE results, batch-corrected expression profiles

Nextflow nf-core Salmon DESeq2

GitHub

Fast and efficient preprocessing of scRNA-seq with kallisto | bustools | kb-python

Single-Cell

End-to-end pipeline for scRNA-seq preprocessing using kallisto, bustools, and kb-python — from raw FASTQ files to filtered count matrices ready for downstream analysis with scanpy.

Input: Raw FASTQ files (10x Chromium or similar)

Output: Filtered cell × gene count matrix, QC metrics

Python kallisto bustools kb-python scanpy

Google Colab

A Practical Guide for Single-Cell Data Analysis with scverse Ecosystem

Single-Cell

Comprehensive walkthrough of the scverse single-cell analysis workflow — covering quality control, normalization, dimensionality reduction, clustering, marker gene identification, and cell type annotation using scanpy and related tools.

Input: Count matrix (AnnData .h5ad or 10x format)

Output: Annotated cell clusters, UMAP embeddings, marker genes, cell type labels

Python scanpy scvi-tools AnnData

Google Colab

Predicting Protein Structures with ColabFold and AlphaFold2 in Google Colab

Protein Structure Prediction

Predict 3D protein structures from amino acid sequences using ColabFold’s accelerated AlphaFold2 pipeline — with MSA generation via MMseqs2 and interactive structure visualization directly in the browser.

Input: Amino acid sequence (FASTA format)

Output: Predicted 3D structures (PDB), confidence scores (pLDDT), PAE plots

Python ColabFold AlphaFold2 MMseqs2 py3Dmol

Google Colab

Boltz2-Notebook: Diffusion-Based Protein–Ligand Structure Prediction & Affinity Analysis

Protein Structure Prediction

Predict protein–ligand complex structures and binding affinities using the Boltz2 diffusion model — enabling rapid in silico docking and interaction analysis without traditional molecular dynamics simulations.

Input: Protein sequence and ligand SMILES/SDF

Output: Predicted complex structures (PDB), binding affinity scores, interaction maps

Python Boltz2 RDKit py3Dmol

Google Colab

AI in Drug Discovery: Molecular Property Prediction and Virtual Screening

Drug Discovery

End-to-end pipeline for AI-driven drug discovery — covering molecular featurization, toxicity prediction, ADMET property modeling, and virtual screening of compound libraries using machine learning and cheminformatics tools.

Input: Compound libraries (SMILES), molecular descriptors

Output: Toxicity predictions, ADMET profiles, ranked hit compounds

Python RDKit DeepChem scikit-learn

Google Colab

AI in Drug Discovery: In Silico Toxicology Modeling

Drug Discovery

Build machine learning models to predict compound toxicity from molecular structure — covering molecular fingerprint generation, toxicity endpoint classification, structure–activity relationship analysis, and model interpretation for safety assessment in early-stage drug discovery.

Input: Chemical compounds (SMILES), toxicity endpoint labels

Output: Toxicity classification models, SAR insights, safety predictions

Python RDKit scikit-learn Mordred

Google Colab