Bowtie2 mystery problem: empty fasta file

[susanklein]

Hi,

until recently I was using Bowtie2 with no problems. Today I am trying to build a new index e.g.

bowtie2-build test.fna test

the fasta file was some plasmid sequence but kept giving me the below error after a few mintues 'stalled':


**********************************************
fchr[A]: 0
fchr[C]: 51813
fchr[G]: 65001
fchr[T]: 84678
fchr[$]: 126810
Exiting Ebwt::buildToDisk()
Returning from initFromVector
Wrote 4236774 bytes to primary EBWT file: p.1.bt2
Wrote 31708 bytes to secondary EBWT file: p.2.bt2
Re-opening _in1 and _in2 as input streams
Returning from Ebwt constructor
Headers:
len: 126810
bwtLen: 126811
sz: 31703
bwtSz: 31703
lineRate: 6
offRate: 4
offMask: 0xfffffff0
ftabChars: 10
eftabLen: 20
eftabSz: 80
ftabLen: 1048577
ftabSz: 4194308
offsLen: 7926
offsSz: 31704
lineSz: 64
sideSz: 64
sideBwtSz: 48
sideBwtLen: 192
numSides: 661
numLines: 661
ebwtTotLen: 42304
ebwtTotSz: 42304
color: 0
reverse: 0
Total time for call to driver() for forward index: 00:00:00
Warning: Empty fasta file: 'test.fna'
Warning: All fasta inputs were empty
Total time for backward call to driver() for mirror index: 00:00:41
Error: Encountered internal Bowtie 2 exception (#1)
Command: bowtie2-build --wrapper basic-0 test.fna p
Deleting "p.3.bt2" file written during aborted indexing attempt.
Deleting "p.4.bt2" file written during aborted indexing attempt.
Deleting "p.1.bt2" file written during aborted indexing attempt.
Deleting "p.2.bt2" file written during aborted indexing attempt.

******************************************************

I experimented with the file size, and irrespective of the actual sequence used
my limit is 126811 bases.
Any longer and the program gives the 'empty fasta file' error. Anyone have any clue what is going on?? previously I have had no problems iwth much bigger files on the same machine (80GB RAM).

Thanks,

S

[dpryan]

Try redownloading bowtie2.

[susanklein]

Try redownloading bowtie2.

yes, that was my next move. thanks.

[susanklein]

I just installed the latest version, 2.2.4 from source.

I still get exactly the same file size error... so this appears to really be another file size bug in bowtie2?

Does anyone know how I can fix this. I really need to build an index with this (pseudo) genome.

Btw, bowtie1 gives the same error. So does every other version I have tried a few 10s of bp smaller and the file works. and a few hundred larger and it works again, but the range I need, 204308 bp, no dice.

Thanks for any help.. getting desperate.

S.

[susanklein]

Ok, I thought of a solution.. I've just duplicate the whole sequence in tandem and its worked. I only want to use the reference to filter out reads (keeping the unmapped reads) so this should work just as well..

S.

[dpryan]

Can you post the reference that's giving you problems somewhere? Then one of us can try to reproduce it and see if this actually is a bug or is just some local computer problem.

[susanklein]

ok, I'll post it next chance I get. Thanks.

[susanklein]

Hi, my file is 203Kb and the forum limit is 19.5 Kb, so I can't upload it

[dpryan]

google drive, dropbox, there are plenty of options.

[susanklein]

ok.. here's the reference file. Just tried it again. Same error:

Getting block 8 of 8
Reserving size (38308) for bucket
Calculating Z arrays
Calculating Z arrays time: 00:00:00
Entering block accumulator loop:
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Block accumulator loop time: 00:00:00
Sorting block of length 18959
(Using difference cover)
Sorting block time: 00:00:00
Returning block of 18960
Exited Ebwt loop
fchr[A]: 0
fchr[C]: 83161
fchr[G]: 104600
fchr[T]: 136354
fchr[$]: 204308
Exiting Ebwt::buildToDisk()
Returning from initFromVector
Wrote 4262634 bytes to primary EBWT file: plasmids.1.bt2
Wrote 51084 bytes to secondary EBWT file: plasmids.2.bt2
Re-opening _in1 and _in2 as input streams
Returning from Ebwt constructor
Headers:
len: 204308
bwtLen: 204309
sz: 51077
bwtSz: 51078
lineRate: 6
offRate: 4
offMask: 0xfffffff0
ftabChars: 10
eftabLen: 20
eftabSz: 80
ftabLen: 1048577
ftabSz: 4194308
offsLen: 12770
offsSz: 51080
lineSz: 64
sideSz: 64
sideBwtSz: 48
sideBwtLen: 192
numSides: 1065
numLines: 1065
ebwtTotLen: 68160
ebwtTotSz: 68160
color: 0
reverse: 0
Total time for call to driver() for forward index: 00:00:00
Warning: Empty fasta file: 'plasmids.fna'
Warning: All fasta inputs were empty
Total time for backward call to driver() for mirror index: 00:00:00
Error: Encountered internal Bowtie 2 exception (#1)
Command: bowtie2-build --wrapper basic-0 plasmids.fna plasmids
Deleting "plasmids.3.bt2" file written during aborted indexing attempt.
Deleting "plasmids.4.bt2" file written during aborted indexing attempt.
Deleting "plasmids.1.bt2" file written during aborted indexing attempt.
Deleting "plasmids.2.bt2" file written during aborted indexing attempt.

[susanklein]

.. I just had to gzip it for the forum.

[susanklein]

command was simply:

bowtie2-build plasmids.fna plasmids

(after gunzipping the file obviously)

[dpryan]

This is a local computer problem. I had no problem creating the indices.

[susanklein]

Ok, thanks.

but I can't see why I get this problem. My Debian machine has plenty of memory, no other problems and there is no indication of what is going wrong in the log. Bowtie2 starts building then just seems to stop at a point when it decides the input is empy.. Its very frustrating. And why would it only do it for certain files size??

[dpryan]

Are you running out of space on that partition? Aside from that, I haven't a clue.