Samtools flagstat on Tophat generated .bam file

[rg_gis]

Hi,

On running the samtools flagstat command on the accepted_hits.bam file generated by tophat, I get the following output:

13414556 in total
0 QC failure
0 duplicates
13414556 mapped (100.00%)
0 paired in sequencing
0 read1
0 read2
0 properly paired (nan%)
0 with itself and mate mapped
0 singletons (nan%)
0 with mate mapped to a different chr
0 with mate mapped to a different chr (mapQ>=5)


I am not sure why the duplicates are 0, as this run of tophat allws upto 40 mismatches and there are reads which are mapped to several location on the genome.

Could someone tell me if there is an alternate way to find out the number of unique reads and the distribution of multi mapped reads from the bam file?

Thanks you.

[kmcarr]

Samtools flagstat does not inspect the alignments itself when it reports these statistics, it only reads the status flag value associated with each alignment and summarizes that. The flag has a bit which indicates whether or not a read is a duplicate (x0400, or 1024 in decimal). This flag would need to have been set by some other tool which analyzes the S/BAM file for duplicates, e.g. Picard's markDuplicates.

Also, it seems you are misunderstanding what is meant by a duplicate in NGS mapping experiments. A duplicate does not mean a read which maps to multiple location on the genome. Duplicates refer to multiple reads which map to the identical location on the genome.

[rg_gis]

Thank you for the reply.. I did misunderstand the meaning of duplicates..

As far as I have understood, The NH tag of a .sam file is what determines how many locations the read has mapped to ..

[telos]

Any idea why every single read appears as mapped? Is TopHat not outputting unaligned reads to the bam?

[Jon_Keats]

Yes Tophat only output the alignments. You can do a line count of your input fastq file "wc -l sequences.fastq" to get total number of reads. Remember to divide the line by 4, as each read is encoded in four lines (read ID, Read Sequence, read ID, Quality values). The fun thing with single end reads is when you end up with 120% mapped! Love the fact that tophat prints the same read multiple times at different locations if the read maps to multiple locations.

[DavidMatthewsBristol]

Hi all,

I've been worrying about multihit reads as well. I've put up a workflow on Galaxy that attempts to deal with this by separating the multi hit reads form the unique ones - hope it helps.

David

[archana2287]

Hi all,

I've been worrying about multihit reads as well. I've put up a workflow on Galaxy that attempts to deal with this by separating the multi hit reads form the unique ones - hope it helps.

David

please share link (workflow on galaxy ), that separates the multi hit reads form the unique ones.

waiting for reply