Tweet
Hello SeqAnswers!
I am working with strand-specific RNA-sequencing data from a bacteria. I've attached a screenshot that demonstrates my issue.
Background:
I've used bowtie2 to map paired Illumina reads to the genome. I've split each BAM file by the fragment's strand (FR-stranded library using samtools and conditional statements). Below you see that the reads are colored according to the appropriate strand in IGV. The top track (purple) shows fragments from one replicate aligned to the reverse/minus strand. The second track from the top (pink) indicates fragments aligned to the forward/plus strand. The next two tracks are a biological replicate of the same condition.
The annotation tracks (blue) are the existing CDS annotation, the top replicate's cufflinks assembly, and the merged assembly across all conditions.
Problem:
Below you will see that there are no purple fragments aligned to the minus/reverse strand near the central vertical bar (~9800bp) and yet Cufflinks must be assembling pairs in a pair-unaware manner, although I used the fr-firststrand argument. My 3' overhang tolerance is set to 10bp.
Question:
Has anyone else encountered this pair-unaware behavior in bacterial transcriptome assembly? Does anyone have any solid workarounds? Is there a parameter in cufflinks I should be aware of?
Here is my cufflinks code:
I am working with strand-specific RNA-sequencing data from a bacteria. I've attached a screenshot that demonstrates my issue.
Background:
I've used bowtie2 to map paired Illumina reads to the genome. I've split each BAM file by the fragment's strand (FR-stranded library using samtools and conditional statements). Below you see that the reads are colored according to the appropriate strand in IGV. The top track (purple) shows fragments from one replicate aligned to the reverse/minus strand. The second track from the top (pink) indicates fragments aligned to the forward/plus strand. The next two tracks are a biological replicate of the same condition.
The annotation tracks (blue) are the existing CDS annotation, the top replicate's cufflinks assembly, and the merged assembly across all conditions.
Problem:
Below you will see that there are no purple fragments aligned to the minus/reverse strand near the central vertical bar (~9800bp) and yet Cufflinks must be assembling pairs in a pair-unaware manner, although I used the fr-firststrand argument. My 3' overhang tolerance is set to 10bp.
Question:
Has anyone else encountered this pair-unaware behavior in bacterial transcriptome assembly? Does anyone have any solid workarounds? Is there a parameter in cufflinks I should be aware of?
Here is my cufflinks code:
Code:
$cufflinks -o $CUFFLINKS/${f} -p $CORES -b $REFFASTA -u -M $MASK -m 520 -s 49 -L $f -I 1 --min-intron-length 0 --max-multiread-fraction 0.2 --overlap-radius 5 --3-overhang-tolerance 10 --library-type fr-firststrand -g $REFERENCE --min-frags-per-transfrag 15 $INDIR/$f.3.bam