Bioinformatics Toolkit

The problem

Mapping raw biological data to the correct analysis pipeline is error-prone. Selecting the wrong tool for sequence alignment, variant calling, or RNA-seq quantification leads to downstream failures and invalid biological conclusions.

What it does

• Routes analysis requests into five specialized lanes: sequence analysis, genomics, transcriptomics, variant calling, and protein structure.
• Identifies the correct tool from a registry of 400+ bioinformatics utilities including BWA, SAMtools, GATK, and DESeq2.
• Enforces file contracts by validating required inputs like paired-end FASTQ, BAM, VCF, and GTF annotations.
• Provides canonical command patterns for processing, such as read trimming with fastp or protein folding with ColabFold.
• Defines quality gates, including Q30 thresholds, mapping percentages, and Ti/Tv ratios for VCF validation.

Frameworks & tools

BioPython, BWA, SAMtools, BCFTools, GATK, STAR, Salmon, DESeq2, Snakemake, and Nextflow.

Why this beats prompting it yourself

General LLMs often hallucinate command flags or suggest outdated tools for genomic workflows. This skill implements specific bioinformatics best practices, such as mandatory BAM indexing and pLDDT confidence reporting for protein models, ensuring pipelines are scientifically sound.

Use cases

• Generating shell scripts for raw FASTQ quality control and alignment.
• Building differential expression workflows using Salmon and DESeq2.
• Identifying germline mutations from sorted BAM files using BCFTools or GATK.
• Predicting protein structures from FASTA sequences with confidence metrics.

Known limitations

Not for clinical variant interpretation. Requires adaptation for specific HPC schedulers like SLURM or LSF. Does not support image-based microscopy assays.