Does gffread extract assembled transcript?

[xmubingo]

I have read thread 'extract assembled transcripts' http://seqanswers.com/forums/showthread.php?t=5116.

But i am not clear whether gffread can extract assembled transcript now? I mean I need extract assembled transcripts but not sequences extracted from genome.

The manual says that "The gffread utility can be used to generate a FASTA file with the DNA sequences for all transcripts in a GFF file. For this operation a fasta file with the genomic sequences have to be provided as well. For example, one might want to extract the sequence of all transfrags assembled from a Cufflinks assembly session. " http://cufflinks.cbcb.umd.edu/gff.html

thanks!!

[dpryan]

The point of the GFF/GTF file is to give information regarding the structure of genes or other features. gffread will extract and combine the sequence of those, regardless of what their nature is. The end of the thread you linked to wanted something slightly different. What they wanted were the sequence for those features in their samples, rather than the sequence for their features in the reference genome. If you want that too (it's unclear from what you wrote), then just call SNPs (careful, you're using RNAseq data!), convert the genome such that it includes the SNPs present in your samples (including indels is much more complicated, but I assume that someone put together a program to do that, though there would be MANY edge cases), and then run gffread with that.

[xmubingo]

The point of the GFF/GTF file is to give information regarding the structure of genes or other features. gffread will extract and combine the sequence of those, regardless of what their nature is. The end of the thread you linked to wanted something slightly different. What they wanted were the sequence for those features in their samples, rather than the sequence for their features in the reference genome. If you want that too (it's unclear from what you wrote), then just call SNPs (careful, you're using RNAseq data!), convert the genome such that it includes the SNPs present in your samples (including indels is much more complicated, but I assume that someone put together a program to do that, though there would be MANY edge cases), and then run gffread with that.

Thanks for your reply!! yes, I use "tophat/cufflinks" to assembled the reads into transcripts. Then, i want to extract the sequences of assembled transcripts in my sample. But, i afraid gffread extract sequence of transcripts from reference genome but not from RNA-seq data. So, your suggestion is to convert reference sequence into a new reference sequence that includes SNPs and indels, and then use gffread get the sequences of transcripts. Yes?

pkerrwall posted a solution similar to yours in 19th floor of thread http://seqanswers.com/forums/showthread.php?t=5116. I paste his solution here:

Here is the process that I am using:

samtools mpileup -uf ref.fa accepted_hits.bam | bcftools view -cg - | vcfutils.pl vcf2fq | fq2fa.pl > new_ref.fa
gffread -w transcripts.fa -g new_ref.fa transcripts.gtf

where fq2fa.pl is a bioperl script to convert from fq to fasta

I also have an email into the cufflinks developers to see if there is a way that the gffread utility can be enhanced to get the consensus sequence from the bam file and not the reference genome.

[dpryan]

I wouldn't hold my breath that they'll add that functionality to gffread, that's pretty far out of its purview. So yes, you'll need to make a consensus genome with samtools and then use that with gffread.

[salzberg]

Yes, gffread can extract transcripts, and do many other things. Geo Pertea has now created a standalone package and a website for gffread and other utilities, here:

https://ccb.jhu.edu/software/stringtie/gff.shtml

and this site also points to the Github repo.