Processing SOLiD data from SRA using Tophat

[Helical]

Hello all,

I'm attempting to run Tophat on SOLiD data from an SRA file and running into problems with the fastq file formatting.

After running fastq-dump on the SRA file, I get the following format:

@SRR1119927.1 solid309_20110721_FRAG_BC_yadegari_1_55_1170 length=50
T000002201013000130000000.01...20...2....2.....2...
+SRR1119927.1 solid309_20110721_FRAG_BC_yadegari_1_55_1170 length=50
!+,0,,/'*&/)&&)2%&+2.0%37!7%!!!1%!!!%!!!!5!!!!!5!!!

Executing Tophat like this:

tophat -C -o output --bowtie1 ColorIndex SRR.fastq

Results in the following error:

Error running bowtie:
Too few quality values for read: 2899T33
are you sure this is a FASTQ-int file?

I researched this error and found that the problem may be I need to use the --quals option and provide a separate quality file. So, I split the fastq file into two separate files:

@SRR1119927.1 solid309_20110721_FRAG_BC_yadegari_1_55_1170 length=50
T000002201013000130000000.01...20...2....2.....2...

+SRR1119927.1 solid309_20110721_FRAG_BC_yadegari_1_55_1170 length=50
!+,0,,/'*&/)&&)2%&+2.0%37!7%!!!1%!!!%!!!!5!!!!!5!!!

And ran:

tophat -C --quals -o output --bowtie1 ColorIndex SRR.fastq SRR_qual.fastq

That generates the following error:

Error encountered parsing file SRR.fastq:
Premature end of file (missing quality values for SRR1119927.1 solid309_20110721_FRAG_BC_yadegari_1_55_1170 length=50)

I can't find any information on how to properly format the base and quality files when they are separated so that Tophat can read them. Is this my problem? Or something else?

<EDIT>

I properly formatted the two split files into proper FASTA:

>SRR1119927.1 solid309_20110721_FRAG_BC_yadegari_1_55_1170 length=50
T000002201013000130000000.01...20...2....2.....2...

>SRR1119927.1 solid309_20110721_FRAG_BC_yadegari_1_55_1170 length=50
!+,0,,/'*&/)&&)2%&+2.0%37!7%!!!1%!!!%!!!!5!!!!!5!!!

But now get the following error:

Error running 'prep_reads'
Error: beginning of quality values record not found! (!'/,<&.&&*'%1*%.2(%&20%'&!')!!!%&!!!1!!!!1!!!!!%!!!)

[mastal]

TopHat is probably expecting the data to be in 2 files, .csfasta and .qual.

I think there should be a command 'abi-dump', instead of fastq-dump,

that will produce the file formats that you need.

[mbblack]

Did you use fastq-dump, or abi-dump to generate your original files? If the SRA submission was actually in color space reads, then you should use "abi-dump" NOT fastq-dump with the SRA toolkit. The abi-dump command will actually give you matched csfasta/csqual files.