SAM Specification The workbench aims to import and export SAM and BAM files according to the v1.4-r962 version of the SAM specification (see http://samtools.sourceforge.net/SAM1.pdf). This appendix describes how the workbench exports SAM and BAM files along with known limitations.
SAM and BAM Export - General notes
The SAM exporter writes unsorted SAM and BAM files.
If the reference name contains spaces, the spaces are removed. Each occurrence of '=' (equals sign) and '@' (at sign) in a reference name is replaced by an '_' (underscore).
The SAM importer and exporter support the ID, SM, PI and PL read group tags. All other read group tags are ignored.
The BAM exporter can also output additional annotations added by tools provided by plugins, and where that is the case, further details are provided in the plugin manual.
SAM Alignment Section A few remarks on the exported alignment section:
- Unmapped reads are not exported.
- If pairs are not on the same contig, the mates will be exported as single reads.
- Multi segment mappings will be imported as a paired data set.
- If a read name contains spaces, the spaces are replaced by an underscore '_'.
- The exported CIGAR string uses 'M' to indicate match or mismatch and does not use '=' (equals sign) or 'X'.
- CLC software does not support or record mapping quality for read mappings. To fulfill the requirement in the BAM format specifications that a read mapping quality is recorded for all mapped reads, the values 0 and 60 are used when mappings are exported from the Workbench. The value 60 is given to reads that mapped uniquely. The value 0 is given to reads that could map equally well to other locations besides the one being reported in the BAM file.
This scoring system is based on a recommendation provided in the SAM FAQ:
Optional fields in the alignment section The following is true for the export of optional fields:
- The NH tag is exported.
- The NM tag is not exported.
- The workbench exports color space information in the CS tag.
- The colors of a right mate are incorrect since the colors of a paired read are stored as a single color string.
- For hard clipped sequence reads, the color space is incorrect, since the color space string is not hard clipped.
- SAM files contain sequence quality score and color quality
scores. The workbench only have color quality scores and these are
stored and exported as sequence quality scores.