multireads in BAM file?

[dglemay]

Hello,

Given X total reads in a FASTQ file, when the reads are aligned (by TopHat), there are greater than X reads mapped in the BAM file (according to samtools flagstats). How can this be?

In flagstats, is a single read counted more than once if it maps to more than one location?

Thanks,
Danielle

[kopi-o]

Yes. Flagstat shows the number of alignments, not the number of reads.

[dglemay]

Thank you, kopi-o!

Any suggestion on how to determine the percentage of mapped and unmapped reads from a BAM file? (I am totally new to samtools)

Thanks!

[kopi-o]

I usually count the number of distinct reads in the bam file by looking at the first column (where the read ID is) and counting the number of unique names using something like

samtools view out.bam | cut -f1 | sort | uniq -c

There are probably more SAM-like ways.