View Single Post
Old 04-11-2011, 05:45 AM   #1
Location: Land of ice and snow

Join Date: Oct 2009
Posts: 10
Default samtools vs picard index

I just stumbled upon something weird. I have a library specific bam file, with several illumina lanes of PE data merged. The file is coordinate sorted and I wanted to index it, like I have past several years, with

samtools index <bamfile>
The original bam file is 7G and I could not believe my eyes when the resulting *.bai file was 14G, so I repeated a few times with the same results. I then decided to test Picard:

java -jar BuildBamIndex.jar
Which behaves and gives me an index file of 7.6M.

samtools idxstats
Yields the same results for both (although it takes much longer with the larger index file, presumably just I/O)

Anyone seen this weird behaviour before?
pallo is offline   Reply With Quote