I am looking into parameters of mpileup, and have a few questions. I'll start with the -C parameter which is recommended to be set to 50 for bwa alignments. According to the docs, the C parameter is a coefficient for downgrading mapping quality of reads. The new mapping quality is set to sqrt((INT-q)/INT)*INT, where INT is the value of -C parameter and q is a phred-scaled probability of the read being generated from the mapped position.
Is the q in the equation equal to the MQ scores in the bam file (mapping quality in my understanding)? Does it never get higher than 50 (MQ does)? If it does, how is the sqrt calculated for a negative number (if default -C50 is used)?
The second part is interference of the -C parameter with the -q parameter (they are both related to mapping quality). The q parameter filters out alignments with mapping scores lower than the given value. I noticed that the value given to the -q parameter has effect on the variants called by mpileup only when -C is not used. Could it be that the -C option changes mapping qualities so much that most/none reads are filtered by -q ? Or does the -C parameter simply override -q?
Any input appreciated.
Is the q in the equation equal to the MQ scores in the bam file (mapping quality in my understanding)? Does it never get higher than 50 (MQ does)? If it does, how is the sqrt calculated for a negative number (if default -C50 is used)?
The second part is interference of the -C parameter with the -q parameter (they are both related to mapping quality). The q parameter filters out alignments with mapping scores lower than the given value. I noticed that the value given to the -q parameter has effect on the variants called by mpileup only when -C is not used. Could it be that the -C option changes mapping qualities so much that most/none reads are filtered by -q ? Or does the -C parameter simply override -q?
Any input appreciated.
Comment