Error running Tophat2 + Bowtie 1 on old SOLiD4 RNAseq SE 50bp data

[aromanowski]

Dear All,
I have downloaded some data from SRA (SRP011410), which were on an ABI SOLiD4 (50bp RNAseq reads).

I am trying to align the reads to the Ensembl TAIR10 reference genome with bowtie 1 and I get the following error:

[2018-02-13 18:59:36] Beginning TopHat run (v2.1.1)
-----------------------------------------------
[2018-02-13 18:59:36] Checking for Bowtie
		  Bowtie version:	 1.2.2.0
[2018-02-13 18:59:36] Checking for Bowtie index files (genome)..
[2018-02-13 18:59:36] Checking for reference FASTA file
[2018-02-13 18:59:36] Generating SAM header for genome-color
[2018-02-13 18:59:37] Preparing reads
	 left reads: min. length=50, max. length=50, 43369633 kept reads (234454 discarded)
[2018-02-13 19:08:20] Mapping left_kept_reads to genome genome-color with Bowtie 
	[FAILED]
Error running bowtie:
Reads file contained a pattern with more than 1024 quality values.
Please truncate reads and quality values and and re-run Bowtie
terminate called after throwing an instance of 'int'

In order to get to that stage, I:
1) I converted the SRA data to .csfata and .quals files using the abi-dump (v2.8.2) command from the sra toolkit.
2) I used the Ensembl TAIR10 genome to build the colorspace indexes for bowtie 1, using the command:

bowtie-build -C genome.fa

3) I ran the following code for tophat2:

tophat2 -p 8 -I 5000 --bowtie1 --color --quals --library-type=fr-secondstrand -o <output_dir> genome-color <csfasta file> <quals file>

I appreciate any help that you can give me to solve this issue!!!

Thank you very much!

Best regards,
Andres

[gringer]

I expect that you should be putting csfastq files into bowtie, rather than csfasta + qual.

But you'd be a lot better off leaving SOLiD data alone. Replicate the experiment using full-length cDNA sequencing on a MinION; you'll get more data that is much easier to reliably match to isoforms, and probably faster and cheaper than it'll take to sort through the minefield of SOLiD bioinformatics.

[aromanowski]

I expect that you should be putting csfastq files into bowtie, rather than csfasta + qual.

Thank you for your advice! I tried using the fastq (generated with fastq-dump instead of abi-dump) and got the same error:

tophat2 -p 8 -I 5000 --bowtie1 --color --library-type=fr-secondstrand -o Col-0_WL_01_thout genome-color SRR444071.fastq 

[2018-02-19 11:55:17] Beginning TopHat run (v2.1.1)
-----------------------------------------------
[2018-02-19 11:55:17] Checking for Bowtie
		  Bowtie version:	 1.2.2.0
[2018-02-19 11:55:17] Checking for Bowtie index files (genome)..
[2018-02-19 11:55:17] Checking for reference FASTA file
[2018-02-19 11:55:17] Generating SAM header for genome-color
[2018-02-19 11:55:17] Preparing reads
	 left reads: min. length=50, max. length=50, 40721267 kept reads (225627 discarded)
[2018-02-19 12:01:53] Mapping left_kept_reads to genome genome-color with Bowtie 
	[FAILED]
Error running bowtie:
Reads file contained a pattern with more than 1024 quality values.
Please truncate reads and quality values and and re-run Bowtie
terminate called after throwing an instance of 'int'

Maybe I should quality filter these reads before starting the mapping? I could definitely try that...

I also tried using the STAR aligner with this same .fastq file and got 0 mapped reads.

Your suggestions sound more and more enticing every minute that goes by:

But you'd be a lot better off leaving SOLiD data alone. Replicate the experiment using full-length cDNA sequencing on a MinION; you'll get more data that is much easier to reliably match to isoforms, and probably faster and cheaper than it'll take to sort through the minefield of SOLiD bioinformatics.

I got a similar dataset (not exactly same tissue, but same genotypes and conditions) from another group that used Illumina, so will try to map those ones with Tophat2 + Bowtie2.

Anyway, it would have been nice to solve the SOLID issue just for the fun of it. However, time is of the essence and I better get the data instead of satisfying this (now personal) problem! =)

Thank you,
Andres

[gringer]

Something else to try: get rid of reads that have '.' in their sequence:

http://seqanswers.com/forums/showthread.php?t=6297