Discrepancy Between Aligned Read And Displayed Alignment In Igv

Question

[terhorst]

I'm puzzled by a seeming discrepancy between what is displayed in IGV, and what is actually contained in my BAM file. Here is the IGV display:

IGV tracks

The highlighted column is at chr1:1012318. The second read in the pileup aligns with edit distance zero starting at this position. However, here is what this read actually looks like in the BAM file:

HSQ1009_86:8:2105:11117:123998  147     chr1    1012318 0       39S61=  =       1011329 -1050   GGGACTCCGTGGGGGGAGGCGGAGGCTATGGGGACTCCGTGGGGGGAGGCTGAGGCTACGGGGACTCCGTGGGGGGAGGCTGAGGCTACGGGGACTCCGT    #######################################@@;8DDEC@@;7C>@;;:-'CGG@5IIIHHGJGGJJJIJIJJJJJJJJHHHHHFDFFFCCB

As expected, the read has a start alignment position of 1012318. However, if you look at the actual sequence, there are nine bases (GGGACTCCG) which precede what is actually displayed in IGV! I'm confused because I thought the alignment position always marked where the start of the read aligns. Can someone explain what is going on? samtools tview shows the same thing.

Answer 1

[Pierre Lindenbaum]

• the mapping quality of your alignment is '0'...
• The cigar string (39S61=) says that 39 are bases soft-clipped (see http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2723002/ ), the start position of the sam-record should begin after those 39 bases. I've never seen the '=' in a cigar string, I suppose it is just a match..

Answer 2

[terhorst]

Thanks for your reply. I see now my mistake: I was confused because the soft clipping specifies 39 bases to be trimmed, but ostensibly only 9 were being trimmed. What did not notice is that the same TGGGGGGAGGCGG occurs later on in the read, after the soft clipping. Thanks!

P.S. To clear up the points you raised for whoever else might be reading this, the mapper I am using is new and somewhat experimental, so mapq may be incorrect or null. As for the '=' in CIGAR, that is actually to spec, but I think bwa doesn't use it so most people haven't seen it before.

[Steve Lianoglou]

FYI: This should be added as a comment to Pierre's answer, and not an answer itself.