TopHat output

[stonerc]

Hi,
I ran TopHat on my RNA-Seq data and I am trying to get a sense of the total number of reads that aligned to anywhere in my reference sequence. Does the accepted_hits.sam file that is part of the TopHat output also report the alignments from the de novo junction discovery? In other words, in order to obtain the total number of mapped reads, can I just use the alignment statistics from the SAM file, or do I also need to add to that the junction counts from the junctions.bed file (5th column reporting the number of alignments spanning each junction)?
Thanks.

[Jon_Keats]

The SAM file contains all the alignments. Just be careful as reads that map to multiple locations get printed at each location so you can get 100+% aligment. But if you filter by read ID you can get the actual percentage of aligned reads.

[lazap]

I'm just starting out with TopHat and it would be great if I could get an idea of how many reads aligned for comparing different options etc.

Can anyone suggest some code for filtering a SAM/BAM file so there is only one alignment for each read and then counting them?

I have tried using
awk '{if (a[$1]++ == 0) print $0; }'
and counting the number of lines in the resulting file but I always get half what is in the SAM/BAM file which is still way above the number of reads I had to start with.

[jgarbe]

Here's what you need to run:
samtools view accepted_hits.bam | cut -f 1 | sort | uniq | wc -l

From http://vallandingham.me/RNA_seq_diff...xpression.html