Variant Annotation Pipeline

Overview

This repository contains an automated annotation pipeline for processing human VCF files using multiple annotation tools, listed below:

• ANNOVAR
• SnpEff
• Ensembl VEP
• BCFtools/csq
• FATHMM-MKL

The Python pipeline (five_tools.py) was developed to evaluate variant annotations from curated ClinVar and OMIM datasets, especially for single-nucleotide variants (SNVs), which are related to Mendelian and complex disorders. The personally curated VCF files can be found in this GitHub page above, named Clinvar_SNVs.vcf.gz and OMIM_SNVs.vcf.gz. The files contain a total of 3,226,691 SNVs, separated relative to their benchmark, OMIM and ClinVar databases, to assess the software tools' accuracy and consistency.

---

Dependencies

Please make sure that the following tools are installed and accessible:

• perl (for ANNOVAR)
• python ≥ 3.8
• java ≥ 8 (for SnpEff)
• bcftools ≥ 1.9
• samtools
• tabix
• ANNOVAR
• SnpEff
• Ensembl VEP
• FATHMM-MKL

---

Setting up your Environment

It is recommended to run this pipeline in a conda environment (especially for VEP) to manage tool dependencies and avoid any errors and conflicts:

conda create -n variant_env python=3.10
conda activate variant_env
conda install -c bioconda ensembl-vep bcftools samtools

Also, please make sure that the following tools are downloaded manually and placed in your home directory:

~/annovar/
~/snpEff/
~/fathmm-MKL/

---

Directory Structure

Your directory should be structured like this:

/home/user/
├── annovar/
├── snpEff/
├── fathmm-MKL/
├── input.vcf
└── five_tools.py

Usage

To run the pipeline, simply call the following from the command line:

python five_tools.py /full/path/to/input.vcf

This will sequentially run:

1. ANNOVAR annotation
2. SnpEff annotation
3. Ensembl VEP annotation
4. BCFtools/csq annotation
5. FATHMM-MKL prediction

---

Output

The following output files will be generated in your working directory, according to the software tools, with their relative variant annotations:

• output_annovar.hg38_multianno.txt
• output_snpeff.vcf
• output_vep.vcf
• output_bcftools_csq.vcf
• output_predictions.txt (for FATHMM-MKL)

---

License

This project is for academic and research use only. Please cite the original tools (ANNOVAR, SnpEff, VEP, etc.) in any publications.

---

Contact

Developed by: Martina Debnath | MSc Genetics and Multiomics in Medicine | UCL

Thank you for using my multi-annotator tool <3

Feel free to reach out for collaborations.

GitHub: @marti-dotcom

Email: [email protected]

---