When running Bowtie in Chrysalis, unable to decipher reads

[Simonli]

Hi Guys,
I`m a RNAseq beginner, I`m using ConDeTri to trim and filter the reads, and inchworm is running good, generate a both.fa file, however when I running the step two of Trinity, it shows my reads could not be deciphered as being /1 or /2 of a PE fragment.
I googled a few related topics but nobody give out a solid answer or solution, anybody has experienced this before?

Thanks,
Simon

---- Chrysalis ----
-------------------

mkdir: cannot create directory `/local_scratch/pbs.6227496.pbs01/trinityoutput/chrysalis': File exists
###########################################################
## Running Bowtie to map fragment ends to Inchworm Contigs
## to use pairing info for Chrysalis' clustering
###############################################################

CMD: ln -sf /local_scratch/pbs.6227496.pbs01/trinityoutput/inchworm.K25.L25.DS.fa target.fa
CMD: bowtie-build -q target.fa target
CMD: bowtie -a -m 20 --best --strata --threads 16 --quiet --chunkmbs 512 -q -S --sam-nohead -f target both.fa > bowtie.out
CMD: /local_scratch/pbs.6227496.pbs01/trinityrnaseq_r2013_08_14/Chrysalis//../trinity-plugins/coreutils/bin/sort --parallel=16 -T . -$
CMD: rm bowtie.out
CMD: /local_scratch/pbs.6227496.pbs01/trinityrnaseq_r2013_08_14/Chrysalis//../util/scaffold_iworm_contigs.pl bowtie.nameSorted.sam /l$
warning, ignoring read: HISEQ:190:C2L5LACXX:6:1101:10000:100038 since cannot decipher if /1 or /2 of a pair.
warning, ignoring read: HISEQ:190:C2L5LACXX:6:1101:10000:100038 since cannot decipher if /1 or /2 of a pair.
warning, ignoring read: HISEQ:190:C2L5LACXX:6:1101:10000:100652 since cannot decipher if /1 or /2 of a pair.
warning, ignoring read: HISEQ:190:C2L5LACXX:6:1101:10000:100652 since cannot decipher if /1 or /2 of a pair.
warning, ignoring read: HISEQ:190:C2L5LACXX:6:1101:10000:11363 since cannot decipher if /1 or /2 of a pair.
warning, ignoring read: HISEQ:190:C2L5LACXX:6:1101:10000:11363 since cannot decipher if /1 or /2 of a pair.
warning, ignoring read: HISEQ:190:C2L5LACXX:6:1101:10000:11363 since cannot decipher if /1 or /2 of a pair.
warning, ignoring read: HISEQ:190:C2L5LACXX:6:1101:10000:15314 since cannot decipher if /1 or /2 of a pair.
warning, ignoring read: HISEQ:190:C2L5LACXX:6:1101:10000:15314 since cannot decipher if /1 or /2 of a pair.
warning, ignoring read: HISEQ:190:C2L5LACXX:6:1101:10000:21043 since cannot decipher if /1 or /2 of a pair.
warning, ignoring read: HISEQ:190:C2L5LACXX:6:1101:10000:21043 since cannot decipher if /1 or /2 of a pair.
number of read warnings exceeded 10. Turning off warning messages from here out.
ERROR: note there were 326239257 reads that could not be deciphered as being /1 or /2 of a PE fragment. Hopefully, these were SE reads that should have been ignored. Otherwise, please research this further.

[atcghelix]

What do the first few reads in the both.fa file look like?

[Simonli]

I see the problems now, check below, I don`t have the "/1" or "/2" in my both.fa file. How is that happening? Should I check the first step of the inchworm?

>HISEQ:190:C2L5LACXX:7:1101:2055:1989 1:N:0:ATCATG
AACAATTTCATCCTCGAATCCTACCACACGGGCGCTGGCACCGCCTACGCCGTCCTCGCTGAGAACACCACCGAAGCCAACCCTCGCACCTTCTACGTGA
>HISEQ:190:C2L5LACXX:7:1101:3448:1999 1:N:0:ATCACG
GTCGTATGCAAAGGATTCAACCCCACGCATTGTAAAATTGCAAGACGCCAGTTAGTCCTAAAGAGTTTCTTCAATAGGACCGTTACCAGCCAGCTAAGAT
>HISEQ:190:C2L5LACXX:7:1101:3524:1986 1:N:0:ATCACG
CTTGTCCTTCTTCTCTTTCCTCTCTTTGGTTTTATCTGACACAAGATCCTTCTCAATCTTATCGAGACTCGGCCTCTGCTTGAGTTTCTCAGAGTCATTA
>HISEQ:190:C2L5LACXX:7:1101:4204:1996 1:N:0:ATCACG
GTGTTAAAAGGGCTAGAGATCGAAATGTTAGACAAAGCGAAATGGAGCAAAGACTGGCAGATCCCAACCAATCCCAATCACGCCGAGAACAAACTTGGTC
>HISEQ:190:C2L5LACXX:7:1101:4587:2000 1:N:0:ATCACG

[Simonli]

Check the reply above..

[atcghelix]

Sounds like Trinity is not properly recognizing the sequence names so that it can add the /1 and /2. Did you modify the read names after receiving the fastq files?

Of potential interest: http://sourceforge.net/mailarchive/f...tyrnaseq-users

You could manually rename the reads in the initial fastq files so that they say /1 and /2.

[Simonli]

Sounds like Trinity is not properly recognizing the sequence names so that it can add the /1 and /2. Did you modify the read names after receiving the fastq files?

Of potential interest: http://sourceforge.net/mailarchive/f...tyrnaseq-users

You could manually rename the reads in the initial fastq files so that they say /1 and /2.

Thanks atcghelix,

When you say manually rename the reads, do you konw what kind of codes can do that?
I`m new to bioinformatics as well as linux.
Thanks.

[Simonli]

Manually put /1 or /2 in the headers, not work

I have manually put /1 or /2 in the headers, still doesn`t work.....