Cufflinks question

[vivekk]

Hi,

I'm trying to set up a local version of Galaxy with all the NGS tools. I'm using this primarily for miRNA-seq data. So the workflow I'm using right now is:

FASTQ Groomer > Clipper > Map w/ Bowtie for Illumina > Filter SAM (for mapped) > Sort SAM for Cufflinks > Cufflinks

Now, this pipeline is working fine. I'm getting the 3 outputs from Cufflinks with FPKM etc. My question is this:

Is there a way in Cufflinks to generate "sequence-level" RPKM output from the SAM file? For example:

If I'm considering let-7a - my reads contain different versions/isoforms of different lengths:

let-7a --- TGAGGTAGTAGGTTGTATAGTT --- "x" RPKM
let-7a --- TGAGGTAGTAGGTTGTATAGT --- "y" RPKM
let-7a --- TGAGGTAGTAGGTTGTATAGTTT --- "z" RPKM

Can I get an output like the one above?
When I run Cufflinks on my Bowtie mapped SAM file, I get the FPKM counts for the miRNAs where - either all the isoforms mapping to the same miRNA are combined or only the most abundant isoform is considered (I'm not sure which). In any case, no sequences are reported in the Cufflinks output.

By the way, I'm mapping to miRBase hairpins.

Any suggestions (Cufflinks or otherwise) about getting this kind of output would be greatly appreciated.

Thanks!
Vivek

[hsmart]

Dear All,

I am trying to use cufflinks to analyze RNA-seq data from two cell-lines. I used following commands:
cuffcompare -i ~/Cufflink_files.txt -r ~/Homo_sapiens.GRCh37.60.gtf -R -p cell1_cell2 -o ~/cell1_cell2_results.txt
cuffdiff -L cell1,cell2 -p 4 -N --FDR 0.05 -r ~/hg19.fa ~/cell1_cell2_results.combined.gtf ~/cell1_accepted_hits.bam ~/cell2_accepted_hits.bam -o ~/Cuffdiff/
I have few questions regarding output of cufflinks:
(1) None of the .diff output files (gene, cds, isoform, promoters, splicing etc) have gene name or gene id associated with it. How can I generate output with gene names? Do I need to change any parameter in cuffcompare?
(2) Also, the locus region of the gene_exp.diff seems to be quite large (for example about 32 kb and includes cluster of 3 genes). So, how does cufflinks define a gene and boundaries related to it?
(3) Also, the locus region of the isoform_exp.diff seems to be quite large (for example about 10 kb and includes entire gene). So, how does cufflinks define an isoform and boundaries related to it?
(4) What type of statistical method does cufflinks use to calculate uncorrected p-value?
(5) What is the meaning of column 7 and 8 (Reserved with value of 0) in splicing.diff file?
(6) How do you compare FPKM and RPKM in terms of absolute values to consider if the gene is expressed above the background?

I really appreciate your personal help regard these issues.

Thanks,

Rakesh

[honey]

GTF file

Rakesh,

One of the first thing you may like to do is to use correct reference annotation GTF file as mentioned in this post

http://seqanswers.com/forums/showthread.php?t=8247

This will add gene names etc.

and will solve some of your issues. So far as 0 in col 7 and 8 is concerned it is because of formatting of the output and are just reserved columns.

Best

[hsmart]

Hi,

Thank you so much for your help.
It worked perfectly with
awk '{print "chr"$0}' Homo_sapiens.GRCh37.60.gtf | sed 's/chrMT/chrM/g' > hg19.ensembl-for-tophat.gtf

I really appriciate your help,

Best wishes,

Rakesh