Old(?) bowtie file: "missing quality values" error in tophat

[worm_picker]

Hello everyone,

I'm trying to look at some old RNA-seq data that I was able to find on NCBI. The data is available as a bowtie output, and I'm trying to use tophat2 to get transcript data.

It originally looked something like this:

HWI-EAS283:1:1:4:1142#0/1 - chr2 70362272 TANTCNTTCCAAGGCTTCTAACATGATGATACTATTTCCTCG B9%<'%<B2;?ACA*B@/BB@;BCCCBBBC@BBCBBA=CB7B 2 36:G>N,39:C>N
HWI-EAS283:1:1:4:1142#0/1 - chr18 50187254 TANTCNTTCCAAGGCTTCTAACATGATGATACTATTTCCTCG B9%<'%<B2;?ACA*B@/BB@;BCCCBBBC@BBCBBA=CB7B 2 36:G>N,39:C>N

Tophat gave the following error:

Traceback (most recent call last):
File "/opt/local/bin/tophat", line 2346, in <module>
sys.exit(main())
File "/opt/local/bin/tophat", line 2251, in main
params.read_params = check_reads(params.read_params, reads_list)
File "/opt/local/bin/tophat", line 1063, in check_reads
if first_line[0] in "@>":
IndexError: string index out of range


So I figured it must be the lack of an '@' at the beginning of the name of the reads, so I used vim to add an @ to the beginning of every line:


@HWI-EAS283:1:1:4:1142#0/1 - chr2 70362272 TANTCNTTCCAAGGCTTCTAACATGATGATACTATTTCCTCG B9%<'%<B2;?ACA*B@/BB@;BCCCBBBC@BBCBBA=CB7B 2 36:G>N,39:C>N
@HWI-EAS283:1:1:4:1142#0/1 - chr18 50187254 TANTCNTTCCAAGGCTTCTAACATGATGATACTATTTCCTCG B9%<'%<B2;?ACA*B@/BB@;BCCCBBBC@BBCBBA=CB7B 2 36:G>N,39:C>N


Now when I run this, I get the following error, where '###' is the file path, sorry wanted to keep that private :


Error encountered parsing file /#############:
Premature end of file (missing quality values for HWI-EAS283:1:1:4:1142#0/1 - chr70362272 TANTCNTTCCAAGGCTTCTAACATGATGATACTATTTCCTCG B9%<'%<B2;?ACA*B@/BB@;BCCCBBBC@BBCBBA=CB7B 2 36:G>N,39:C>N)


This is the very first line, so it seems to hint at a format error...

I looked at the bowtie manual, and it seems my output differs in one way from the manual's: column 5 should be "read sequence", a '+' or '-' value, not the quality score - meaning in between the read and the quality of my output should be another column with a '+' or '-' value.

Am I missing something here? The bowtie output that I'm downloading looks "mostly" like a bowtie output, but it appears wrong... I tried to see if it was maybe an older format, but I can't find any info on that.

Can anybody help me out?

Thanks in advance!!

-worm_picker

[Brian Bushnell]

I don't recognize the format; maybe it was some old bowtie-specific output. If you want to map that, you should convert into fastq format, like this:
@1
5
+
6

...where 1 is the first field (read name), 5 is the 5th field (bases), and 6 is the 6th field (qualities).

[worm_picker]

Thanks for replying, brian.

It's the same as bowtie output normally, and is already mapped, but is just missing a column (I think). On the GEO accession page it claims to be mapped reads from bowtie.

bowtie should be (according to the manual):
1. name
2. strand
3. "contig"
4. 0-based start on contig
5. read
6. read strand
7. quality
8. mismatches (if any)

[Brian Bushnell]

Well, if you want to map it with Tophat, I think you'll have to convert it to fastq first, even if it is (almost) in an old Bowtie format. I doubt you will find any downstream RNA-seq analysis tools that accept those mappings; they generally require sam or bam.