Bioscope multi-hits

[eoh001]

Hi
I found a strange mapping result by Bioscope.
I analyzed the paried end reads from a whole transcriptome, and checked the mapping results.
Threre were a number of reads which were both properly mapped and unmapped at the same time.
For instance, a pair of reads (F3 & F5) '504_910_1119' were mapped poperly, however, there were extra alignments that the F3 was unmapped. I set the 'Alignment filter modes' as 'Primary', so the one best alignment was supposed to be selected among multi-hits. Is it a bug of Bioscope? Any one has observed this kind of results?

**********************************************************
504_910_1119 83 chr21 32600530 0 4H46M = 32600531 50 TTTTTTTAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA @IIIIB1=IIEIIIIHIIIIIIIIIIIIIIIIIIIIIII%%IIII; RG:Z:20110812083702784 NH:i:4 CM:i:3 SM:i:4 CQ:Z:;?;:?7%=;?;9:=929?5:<8==955=<790==:15===12==>3.;=< CS:Z:T3000030000000000000000000000000000000000000000000

**********************************************************
504_910_1119 163 chr21 32600531 0 25M10H = 32600530 -50 TTTTTTAAAAAAAAAAAAAAAAAAA IED86:IE//%%5,,%&-+*/=:/0 RG:Z:20110812083702784 NH:i:4 CM:i:2 SM:i:1 CQ:Z:B)=(1&5=)')%1%(%&''%&*4')(%7%%%%5(. CS:Z:G10000030000200003000000003303000020

**********************************************************
504_910_1119 69 * 0 0 * chr21 18218140 0 NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN * RG:Z:20110812083702784 NH:i:4 CQ:Z:;?;:?7%=;?;9:=929?5:<8==955=<790==:15===12==>3.;=< CS:Z:T30000030000000000000000000000000000000000000000000

**********************************************************
504_910_1119 101 * 0 0 * chr21 34903877 0 NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN * RG:Z:20110812083702784 NH:i:4 CQ:Z:;?;:?7%=;?;9:=929?5:<8==955=<790==:15===12==>3.;=<
CS:Z:T30000030000000000000000000000000000000000000000000
**********************************************************

[mbblack]

The "Alignment filter modes", if I'm remembering correctly (I don't have my BioScope manual at hand) refers to what gets taken from the .bam file into the wig file.

To see what is in your .bam file, just look in your alignmentReport.txt file, the line right under the "-----" (should read "Reads mapped, not filtered") is what is actually in your .bam file.

To get just the "unique" or primary reads, you can use the matobam tool and it has a parameter to pull out only the primary (called "unique") reads into a new .bam file for you.

In a pipeline run, you can filter what goes into the merge .bam file using a mapping qv filter cutoff.

I was told that LifeScope has settings to choose what gets dumped into the mapping .bam file, but so far, I can only see settings to alter what gets used from the .bam file for the .wig files and the coverage reports. As far as I can find thus far, LifeScope is the same as BioScope in this regard - the only way to filter what the pipeline puts into the merge .bam file is via a mapping qv cutoff. Otherwise, all non-filtered, mapped reads end up in the .bam file and you then have to pull out those you want to use in your next steps. I may well have missed or not found something in LifeScope yet though.