If I have more than 20 reference sequences, mpileup fails to call the correct consensus after the 21st sequence. This same problem shows up in tview as an incorrect reference sequence at the top with the correct consensus and alignments underneath.
Assume I have 30 reference sequences, ref1, ref2, .... ref30, and Idxstats reports 5 million reads mapped to ref28.
The consensus extracted for ref1~ref20 are good with high quality. However the consensus for ref28 is poor quality, lots of gaps and lowercase regions, and the output is not similar to ref28, it is usually similar to one of the sequences in ref21~ref27.
In tview this problem shows up as ref2x at the top, and the consensus and alignment lines for ref28.
This problem occurs for both bam files generated by Mosaik and Bowtie2. I have been using a workaround by breaking up the references so that each one only has 20.
Assume I have 30 reference sequences, ref1, ref2, .... ref30, and Idxstats reports 5 million reads mapped to ref28.
The consensus extracted for ref1~ref20 are good with high quality. However the consensus for ref28 is poor quality, lots of gaps and lowercase regions, and the output is not similar to ref28, it is usually similar to one of the sequences in ref21~ref27.
In tview this problem shows up as ref2x at the top, and the consensus and alignment lines for ref28.
This problem occurs for both bam files generated by Mosaik and Bowtie2. I have been using a workaround by breaking up the references so that each one only has 20.
