The SNPsplit workflow in more detail
1) sam2bam
Optional. If the supplied file is a SAM file it will first be converted to BAM format (using samtools view).
2) Sorting
Paired-end files might require the input file to be sorted by read ID before continuing with the allele-tagging (Read 1 and Read 2 of a pair are expected to follow each other in the input BAM file). Unless specifically stated, paired-end BAM files will be sorted by read name (using samtools sort -n; output file ending in .sortedByName.bam). For files that already contain R1 and R2 on two consecutive lanes, the sorting step may be skipped using the option --no_sort. Single-end files or Hi-C files generated by HiCUP do not require sorting.
3) Storing SNP positions
SNP positions are read in from the SNP file (which may be GZIP compressed (ending in .gz) or plain text files). The SNP file is expected to be in the following format (tab-delimited):
ID Chr Position SNP value Ref/SNP
18819008 5 48794752 1 C/T
40491905 11 63643453 1 A/G
44326884 12 96627819 1 T/A
Only the information contained in fields 'Chr (Chromosome)', 'Position' and 'Ref/SNP' base are being used for analysis. The genome containing the 'Ref' base is used for 'genome 1 specific reads (G1)', the genome containing the 'SNP' base for 'genome 2 specific reads (G2)'. If reads do not overlap any SNP positions they are considered 'Unassigned (UA)', i.e. they are not informative for one allele or another. In the rare case that a read contains both genome 1- and genome 2-specific base(s), or that the SNP position was deleted the read is regarded as 'Conflicting (CF)'.
It is probably noteworthy that the determination of overlaps correctly handles the CIGAR operations M (match), D (deletion in the read), I (insertion in the read) and N (skipped regions, used for splice mapping by e.g. HISAT2 or STAR). Furthermore, soft-clipping (operation S) is also supported. Other CIGAR operations are currently not supported and cause SNPsplit to die.
4) Tagging report
Upon completion, a small allele-specific tagging report is printed to screen and to a report file (.SNPsplit_report.txt) for archiving purposes.
5) tag2sort
Once the tagging has completed, the tag2sort module reads in the tagged BAM file and sorts it into various sub-files according to their XX:Z: tag. Both single and paired-end files are sorted into the following four categories:
Files with conflicting SNP information (tag XX:Z:CF) are not written out by default.
6) Sorting report
Upon completion, an allele-specific sorting report is printed out on screen and to a report file for archiving purposes (*.SNPsplit_sort.txt). If the sorting was launched by SNPsplit and not run stand-alone (as tag2sort) the sorting report is also appended to the main SNPsplit report (*.SNPsplit_report.txt).