Storing Fastq As Unaligned Bam

Question

[Abhi]

Hey Guys

Just wondering if anyone there is now storing the raw fastq data as unaligned bam files. We are reaching a stage where any space we could potentially save would be beneficial.

Any con of storing fastq as bam? I see some discussion about this on seqanswers: http://seqanswers.com/forums/showthread.php?t=14941

Also any tools that people already have that converts a fastq to bam and vice-a-versa. I know there are few which can do bam to fastq like picard but not sure if fastq to bam is there.

Thanks!

Abhi

[Peter]

This doesn't read like a stackexcahnge question to me - if you want a discussion why not continue on SEQanswers?

[Ian]

This seems relevant to me, handling large NGS data files is an increasing bioinformatic issue.

Answer 1

[Pierre Lindenbaum]

I think a bgzipped fastq file will be always smaller than a BAM file as the BAM file also contains the positions of the alignments.

See also:

http://bioinformatics.oxfordjournals.org/content/early/2011/01/19/bioinformatics.btr014.abstract

Compression of genomic sequences in FASTQ format

and

http://genome.cshlp.org/content/early/2011/01/18/gr.114819.110

Efficient storage of high throughput sequencing data using reference-based compression

Answer 2

[toni]

Hi Abhi,

yes we do this in our team. You can use Picard 'FastqToSam' utility.

Compared to 2 fastq files (plain, not gzipped as suggested by Pierre), an unaligned BAM file allows to save 60°% to 65% of storage space. It is also practical because you can store some useful information (Sample, Library, Run, any useful comments ...) in the header if you want to.