strand specific mayhem for Tophat and Cufflinks

[adrian]

Dear Group,

I never came across errors in Tophat2/Cufflinks analysis and this error is troubling me and I cannot find a proper solution. I am posting in anticipation that I could get some help.

I have a fastq files from RNA-Seq experiment (1x50bp;strand specific dUTP) single-end 50bp reads.

I aligned the files using tophat2.

tophat -p 16 --library-type fr-firststrand -G /HumanGenome/Grch37/Homo_sapiens/Ensembl/GRCh37/Annotation/Archives/archive-2015-07-17-14-31-42/Genes/genes.gtf -o myDir /HumanGenome/Grch37/Homo_sapiens/Ensembl/GRCh37/Sequence/Bowtie2Index/genome fastq1.fastq

I get accepted_hits.bam file along with other files and I do not see any error.


2. Next I ran, cufflinks and I did not get any problems here. I could successfully generate transcripts.gtf file.

cufflinks -p 16 --library-type fr-firststrand -o myDirCL accepted_hits.bam

3. I next did cuff merge using all gtf files in a list - gtfList

cuffmerge -p 16 -g /HumanGenome/Grch37/Homo_sapiens/Ensembl/GRCh37/Annotation/Archives/archive-2015-07-17-14-31-42/Genes/genes.gtf -s /HumanGenome/Grch37/Homo_sapiens/Ensembl/GRCh37/Sequence/Bowtie2Index/genome.fa gtfList

4. Cuffdiff fails:
cuffdiff -p 16 --library-type fr-firststrand -o mycfDiff /HumanGenome/Grch37/Homo_sapiens/Ensembl/GRCh37/Sequence/Bowtie2Index/genome.fa -u merged.gtf samp1a.bam,samp1b.bam samp2a.bam,samp2b.bam


Warning: Could not connect to update server to verify current version. Please check at the Cufflinks website (http://cufflinks.cbcb.umd.edu).
[bam_header_read] EOF marker is absent. The input is probably truncated.
[bam_header_read] invalid BAM binary header (this is not a BAM file).
[16:06:53] Loading reference annotation.


right after saying loading reference annotation, pipeline fails.


I compared this with previous analyses. I never had EOF marker absent with successful cuffdiff runs.

I validated using ValidateSamFile of Picard and I do not see any difference between bam files of successful cuffdiff analysis and failed analysis.

Also Bam file is gzip file and I do see the end of bam file of failed run :

tail accepted_hits.bam | hexdump -C
00000980 54 84 0f 1b e3 be fa f6 50 67 5d b5 92 9c 16 24 |T.......Pg]....$|
00000990 b6 9e 54 68 40 ff 07 3c e5 ef 1d 2f 8e 00 00 1f |..Th@..<.../....|
000009a0 8b 08 04 00 00 00 00 00 ff 06 00 42 43 02 00 1b |...........BC...|
000009b0 00 03 00 00 00 00 00 00 00 00 00 |...........|
000009bb



Can any one suggest what could be wrong here.

I suspect, if I am correct in executing tophat for single-end 50bp, strand-specific reads.
There are 4 fastq files for each sample.

Header of fastq file:

file 1 : @SN930:564:H3Y5YBCXY:1:1101:1228:2226 1:N:0:AGTCAA
file 2: @SN930:565:H3Y5YBCXY:1:1101:1263:2151 1:N:0:AGTCAAC
file 3: @SN930:564:H3Y5YBCXY:2:1101:1433:2070 1:N:0:AGTCAA
file 4: @SN930:565:H3Y5YBCXY:2:1101:1282:2078 1:N:0:AGTCAAC


thanks a lot.
Adrian

[fanli]

You should use fr-firststrand, based on the Tophat manual.

Also see:

https://github.com/griffithlab/rnaseq_tutorial/blob/master/manuscript/supplementary_tables/supplementary_table_5.md

[adrian]

Hi I have a strand specific RNA-Seq experiment. The libraries were prepared based on dUTP and I was told to use forward strand.

I used Tophat to align the fastq files and merged replicates. I get a single bam file and used RseqQC on bam file. I get the following output:

This is PairEnd Data
Fraction of reads failed to determine: 0.0040
Fraction of reads explained by "1++,1--,2+-,2-+": 0.0087
Fraction of reads explained by "1+-,1-+,2++,2--": 0.9873


From RSeqQC documentation,:

1+-,1-+,2++,2–
read1 mapped to ‘+’ strand indicates parental gene on ‘-‘ strand
read1 mapped to ‘-‘ strand indicates parental gene on ‘+’ strand
read2 mapped to ‘+’ strand indicates parental gene on ‘+’ strand
read2 mapped to ‘-‘ strand indicates parental gene on ‘-‘ strand


I understood that the bam files, majority of reads, where read2 mapped to + strand with gene on + strand.

http://seqanswers.com/forums/images/icons/icon14.gif

My question:

How do I set the option --library-type for cufflinks. Cufflinks offfers multiple options for library-type such as:

Supported library types:
ff-firststrand
ff-secondstrand
ff-unstranded
fr-firststrand
fr-secondstrand
fr-unstranded (default)
transfrags

in my case, based on rseq qc, shoudl i choose ff-secondstrand? or fr-firststrand or fr-reversestrand ?

appreciate your help.

Thanks
Adrian.