Bowtie2 cannot read gzip format files

[yumtaoist]

Hey, everyone!

The help document of bowtie2 have said that the files of reads could be gzip'ed.

When I use the paired reads and try the argument ( -1 <m1> -2 <m2>), the bowtie2 can read the gzip'ed files.

However, if I use the gzip'ed unpaired reads and apply the argument (-U <r>), the bowtie2 suggest that "Error: reads file does not look like a FASTQ file". The gzip'ed file can not be recognized.

Does anyone have come up with this problem?

Thank You!

[earonesty]

# HOW to add GUNZIP support to ANY program that doesn't have it:

mkfifo file1.fifo
mkfifo file2.fifo
gunzip file1.gz > file.fifo &
gunzip file2.gz > file.fifo &
bowtie -1 file1.fifo -2 file2.fifo

[yumtaoist]

# HOW to add GUNZIP support to ANY program that doesn't have it:

mkfifo file1.fifo
mkfifo file2.fifo
gunzip file1.gz > file.fifo &
gunzip file2.gz > file.fifo &
bowtie -1 file1.fifo -2 file2.fifo

Thank You Very Much!

This is very useful!

[gringer]

# HOW to add GUNZIP support to ANY program that doesn't have it:

mkfifo file1.fifo
mkfifo file2.fifo
gunzip file1.gz > file.fifo &
gunzip file2.gz > file.fifo &
bowtie -1 file1.fifo -2 file2.fifo

In many cases this will work, but not all cases. This doesn't work for programs that attempt to seek through files (low-memory fastq randomisation being one example that I've recently seen).

[dsenalik]

I needed to add a "-c" to that command, otherwise gunzip will delete the original compressed file, i.e.

mkfifo file1.fifo
mkfifo file2.fifo
gunzip -c file1.gz > file1.fifo &
gunzip -c file2.gz > file2.fifo &

[earonesty]

I needed to add a "-c" to that command, otherwise gunzip will delete the original compressed file, i.e.

mkfifo file1.fifo
mkfifo file2.fifo
gunzip -c file1.gz > file1.fifo &
gunzip -c file2.gz > file2.fifo &

this patch fixes fifo support, and also adds built-in gzip support to bowtie 1.0 (which remains the fastest, fastq->sam program out there)

https://github.com/earonesty/bowtie

[westerman]

FYI, zcat is often a shortcut to 'gunzip -c'

Also under BASH you should be able do something like:

bowtie -1 <( zcat file1.gz ) -2 <( zcat file2.gz )

or in the original posters question

bowtie2 -U <( zcat file.gz )

[earonesty]

actually there's a bug in bowtie ... a race condition that can make it unreliable. you should not use fifo's unless you used the patched version that supports them

https://github.com/earonesty/bowtie