BSfilter is a commandline-interfaceded python3 script that takes at lease one argument and at most 3 arguments. Its purpose is to parse through a fastq file and output a fastq file that contains reads with a specified frequency and/or count of Cytosines per read.
Software:
- python 3.7.3
BSfilter can be run from a unix-based commandline using the following syntax:
python3 BSfilter.py -fq <path/to/fastq/file> [-options]
example:
python3 BSfilter.py -fq ./reads.fastq -mC 0.05
returns fastq file with non-converted reads (reads with high cytosine ratio) removed
-fq <string> path/to/fastq/file.
-mC <float> max frequency of Cytosines allowed per read. Default = 1.0
--count <integer> max number of Cytosones allowed per read. Default = 1000
BSfilter was designed to filter out non-bisulfite-converted transcripts from fastq files. Hyperthermophilic organisms require very harsh bisulfite conversion conditions. These conditions will lead to complete conversion of spike in transcripts, making it impossible to deduce non-conversion rates. One solution is to remove reads from fastq files that have a higher than expected frequency of Cytosines (methylation calls). Although Bisulfite-seq mapping tools allow you to tag reads with a specified cytosine frequency and eliminate them from downstream analysis, BSfilter allows you to optimize methylation threshholds via visual inspection in a genome browser, such as IGV.
The top panel is a genome region that was previously unfiltered. The subsequent panels were subjected to filtering allowing 5%, 3%, and 2% Cytosines per read.