Cufflinks is frozen

Question

[Rox]

Hi everyone !

This is my first time using Cufflinks.

I'm trying to get a fasta corresponding to the RNA-seq from different tissues from D. suzukii. I first used hisat2 in order to align my reads against a reference, here the command I used in Hisat2 :

hisat2 -p 25 --max-intronlen 50000 -x /media/DATAPART1/Documents/suzukii_assembly/annotation/indexes/suz_canu3 -1 '/media/DATAPART1/Documents/suzukii_assembly/annotation/EST_evidences/suz_ant.R1.fastq.gz','/media/DATAPART1/Documents/suzukii_assembly/annotation/EST_evidences/suz_ovi.R1.fastq.gz','/media/DATAPART1/Documents/suzukii_assembly/annotation/EST_evidences/suz_pro.R1.fastq.gz','/media/DATAPART1/Documents/suzukii_assembly/annotation/EST_evidences/suz_tar.R1.fastq.gz' -2 '/media/DATAPART1/Documents/suzukii_assembly/annotation/EST_evidences/suz_ant.R2.fastq.gz','/media/DATAPART1/Documents/suzukii_assembly/annotation/EST_evidences/suz_ovi.R2.fastq.gz','/media/DATAPART1/Documents/suzukii_assembly/annotation/EST_evidences/suz_pro.R2.fastq.gz','/media/DATAPART1/Documents/suzukii_assembly/annotation/EST_evidences/suz_tar.R2.fastq.gz' | samtools view -Sbo suzukii_rnaseq_ibdm.bam -

Then I sorted the bam file using this command :

samtools sort /media/DATAPART1/Documents/suzukii_assembly/annotation/suzukii_rnaseq_ibdm.bam /media/DATAPART1/Documents/suzukii_assembly/annotation/suzukii_rnaseq_ibdm_sorted.bam

Something that bothered me at this step is, first it took a long time but well, and second, the sorted file was shorter than the original file (37.2G -> 26.2G), and at this point I don't know if this was normal, but I tried anyway to launch Cufflinks using this sorted file.

At first, I used this command :

cufflinks --library-type fr-firststrand -o /media/DATAPART1/Documents/suzukii_assembly/annotation/cufflinks/hisat2 -p 25 /media/DATAPART1/Documents/suzukii_assembly/annotation/suzukii_rnaseq_ibdm_sorted.bam

But I get strange errors like this :

BAM record error: found spliced alignment without XS attribute

I read a post on Biostar (don't remember which one), and the problem could be the --library-type, so I tried with this option on :

--library-type fr-firststrand

And then it seems to work for a moment, but it froze at 14% of the sequence (I tried 2 times) :

You are using Cufflinks v2.2.1, which is the most recent release.
[17:08:27] Inspecting reads and determining fragment length distribution.
Processing Locus tig00000383:187849-193055   [***                      ]  14%

Cufflinks didn't stop, it didn't throw any error, it is still doing stuff apparently (I'm watching it with htop), but it run all nigh, and I think that something went wrong...

So, is there something wrong with this particular locus ? Is the sorting step a problem ? Or I've just used the wrong library type ?

Thanks for your help !

Cheers,

Roxane

[WouterDeCoster]

Small comment: It's indeed normal that a sorted bam is smaller than a non sorted bam, because better compression is possible with sorted records. Perhaps someone else will have a more intuitive explanation for this but it's nothing to worry about.

Answer 1

[andrew.j.skelton73]

When using HISAT2, you need to tell the aligner to output "Cufflinks Friendly" alignments using the --dta-cufflinks parameter, see the manual here

[Rox]

Okay ! That's probably why. I should have notice, and I was aware that the same option already exist for tophat2. I'm going to try this then, thanks !

[Rox]

Sadly, I thought that the --dta-cufflinks was the answer, but well, even with rereunning hisat2, I get exactly the same issue, Cufflinks is again frozen at 14% on the same contig and on the same location...

Could it be the --library-type fr-firststrand option ?