bcov

Bcov

Effortlessly query coverage and read depths of genes and exonic variants across dozens of exome probe kits and hundreds of gigabytes of original BAM raw data.

Backend database preloading and serving of the HTTP API and React frontend is done through the same single binary with zero dependencies.

https://pkg.go.dev/github.com/tikz/bcov/cov

https://pkg.go.dev/github.com/tikz/bcov/api

https://pkg.go.dev/github.com/tikz/bcov/db

Local build

Requirements

Go >=1.18 and npm

sudo apt install nodejs golang

Build

git clone https://github.com/tikz/bcov.git
cd bcov

make

./bcov

Bcov
Version:  development                Commit:  unknown

Usage:
  -bam string
        Load a BAM file into the database
  -bams string
        Load multiple BAM files passing a CSV file that in each line contains <bam path>,<capture kit name>
  -delete-bam string
        Delete BAM file from the database with SHA256
  -fetch-exons
        Fetch exons from Ensembl and store in DB
  -fetch-variants
        Fetch SNPs from ClinVar and store in DB
  -help
        Display help
  -kit string
        Capture kit name
  -region string
        Print per position depth of a given range, expressed as <chromosome>:<start>-<end>
  -test-db
        Test database connection
  -web
        Run web server

Database setup

1. Load all GRCh38 exons from Ensembl

./bcov -fetch-exons

Populates the local database fetching from Ensembl core schema all exons in genes that have a valid gene symbol name defined by HGNC and a MANE Select canonical transcript.

2. Load all dbSNP variants in exonic regions

./bcov -fetch-variants

Populates the local database using the latest variant_summary.txt from the NCBI public FTP server.

3. Load a BAM file

./bcov -bam <path> -kit <kitName>

Loads a BAM file and assigns it under the given kit name.

Each BAM file is meant to be loaded only once. A file is deemed unique (by practical purposes) and enforced using the SHA256 sum.

Make sure that the BAM files you intend to load are aligned to GRCh38, otherwise coordinate mappings will be wrong.

Expected performance of traversing a single BAM file is about 1 GB/min using the SQLite engine, a consumer level NVMe SSD, and Ensembl/ClinVar datasets as of 2022. Resulting size in database is about ~1/10 of the original file.

*. Load multiple BAM files

./bcov -bams <csvPath>

Load multiple BAM files passing a CSV file in which each row should contain <path>:<kitName>

bams/P0971.bam, SureSelect Human All Exon V6 (100X)
bams/P0978.bam, SureSelect Human All Exon V7 (200X)

4. Run web server

./bcov -web

Live

https://bcov.qb.fcen.uba.ar