SnpFilt is a package for identifying sequence variants in high-throughput sequencing reads using reference-free assemblies

Installation

SnpFilt has only been tested on a Linux system. Specifically it requires the following programs to be available

• Perl
• R
• gunzip
• zgrep
• cut

It also requires a number of third-party packages

• samtools
• bcftools
• BWA
• SPAdes
• LASTZ
• caTools (R-package)

By default, install.sh will check if any of these can be found in the path If not it will install missing pacakges from the third_party directory The user can direct the install script to other locations by modifying lines 16-20 in install.sh. Set variable to empty string to force re-installation

By default, the SnpFilt package will be installed in a folder called snpfilt which will be created inside the current directory, but the user can change this by modifying line 8 in install.sh

Once file paths are set correctly, SnpFilt can be installed typing

> ./install.sh

To put in snpfilt into the path you can add the snpfilt directory to your $PATH varable or put a symlink to snpfilt/snpfilt.pl in a directory in your path.

Running SnpFilt

To find sequence variants to a given set of reads type

> <snpfiltpath>/snpfilt.pl pipeline <outdir> <reads1.fastq> <reads2.fastq> <refseq.fasta>

or

> perl <snpfiltpath>/snpfilt.pl pipeline <outdir> <reads1.fastq> <reads2.fastq> <refseq.fasta>

For further instructions regarding usage simply type

> snpfilt.pl 

Test data is provided in the examples directory and detailed comments regarding usage is available in the run_example.sh script.

Output Files

SnpFilt creates a directory for each isolate and generates a number of output files in the directory

• contigs.fasta: this the raw assembly generated by SPAdes, which is likely to contain sequence errors
• masked_contigs.fasta: this is assembly with problematic sites masked to 'N'
• flt_contigs.gz: compressed CSV file listing the base and filter code for each base in the assembly. Only sites with flt=0 are reliable. See snpfilt_results_contigs.csv for an explanation on the other filter codes.
• contigs.bcf: mpileup file generated by mapping the read back to contigs.fasta

The files listed above are always created and use only information based on the reads themself. When a reference genome is provided the isolate will be compared with it generate the additional files

• snpfilt_results_fstem.csv: CSV file listing all potential SNPs and the filter code describing their reliability
• aln_fstem.cigarx: tabular file giving positions where assembly aligned to the reference genome