Ribo remover

This is a small utility to remove mammalian ribosomal RNA content from FASTQ files, particularly for RNA-seq. Ribo-remover produces filtered FASTQ files without the rRNA content, using NCBI's blastn utility to identify rRNA, similar to PORT. It supports single end or paired end reads.

Installation

Using python 3.9 or greater, and preferably using a virtual environment, run the following:

pip install git+https://github.com/tgbrooks/ribo_remover

Usage

# For single-end reads
ribo-remover --input-fastqs R1.fastq.gz --output-fastqs filtered.R1.fastq.gz

# For paired-end reads
ribo-remover --input-fastqs R1.fastq.gz R2.fastq.gz --output-fastqs filtered.R1.fastq.gz filtered.R2.fastq.gz

# To also output a file containing ribo content stats as a CSV file
ribo-remover --input-fastqs R1.fastq.gz R2.fastq.gz --output-fastqs filtered.R1.fastq.gz filtered.R2.fastq.gz --stats-file ribostats.csv

Both input and output can be plain fastq files or gzipped files ending in .gz, which will be automatically decompressed (for input files) and compressed (for output files) based off the provided file names.

Multi-threading

You can request multi-threading with the --threads option, but note that this is the number of threads used per fastq file. If you have paired end reads, then two instances of blastn will be run, each using the provided number of threads. In addition, the input/output files may need computational resources to be gzipped. This means that you generally want to have at least 2n+1 cores available when requesting --threads n with paired ends and n+1 cores if using single ends. Performance benefits also drop off after --threads 2.

UMI tags in separate files

Sometimes, unique molecule identifier (UMI) tags are provided as separate files, for example R1.fastq, R2.fastq along side UMI tag files I1.fastq and I2.fastq. In this case, any ribosomal reads will want to be removed from the UMI tag files as well. This should be possible by passing all fastq files (including UMIs) together to --input-fastqs and giving corresponding outputs for --output-fastqs. This will BLAST each UMI tag against the ribosomal database as well, but typical UMI tags are too short to give a significant hit and so should have no real effect on the ribosomal removal. However, this functionality has not been extensively tested.