Tweet
Hi Everyone,
I have some questions on if my results are valid by using cuffdiff.
What I did currently is that I did alignment with bowtie to drosophila transcript-r5.55. Unlike common references, this reference is in the format of following:
>FBtr0086024 type=mRNA; loc=2R:complement(1944862..1947063); ID=FBtr0086024; name=CG7856-RA; dbxref=FlyBase:FBtr0086024,FlyBase_Annotation_IDs:CG7856-RA,REFSEQ:NM_136354; score=11; score_text=Strongly Supported; MD5=7ca9c4371b97aaf7046cf83fefb65eb8; length=2202; parent=FBgn0033056; release=r5.55; species=Dmel;
GGTAAAGTTGCCTTGGCGTCAGTTGGCAGTTTGGGAAAAGCCTACACACT
TAATATTTCGATAGATACACTTATTTCGCAATCGTAGAAGATACCACAAA
TCTCTCTTCCGTAAATTATAAGTATGTCCAAGAGGGTGAGCATCATGTTG
CCCGACGAAATACCTGCGGCTCCGTCAGGCAGCAGGAGGAACCCGATGCC......
So when I got the bam file, the column usually indicating chromosome localization is already filled with FBtr#:XXX-XXX (instead of chr1:XXX-XXX) which could be used as annotation.
I run the bam file in cufflinks to get the GTF file for each file and run cuffmerge to get a merged GTF file for files in the same set of experiment. Then I run cuffdiff with merged GTF file and bam files. I did NOT use any reference GTF files during those steps since the reference for mapping I used is already annotated as I described above. In cuffdiff output file isoform_exp.diff under the column of loci, the value also remained FBtr#:XXX-XXX. It means for each transcript, it is compared in different region. So should I consider if any of the region is differentially expressed, the transcript is differentially expressed? And is this going to affect validity of cuffdiff's calculation since this equals to do assembly first then quantification for differential expression compared to usually it does quantification first then assembly?
Thanks in advance.
I have some questions on if my results are valid by using cuffdiff.
What I did currently is that I did alignment with bowtie to drosophila transcript-r5.55. Unlike common references, this reference is in the format of following:
>FBtr0086024 type=mRNA; loc=2R:complement(1944862..1947063); ID=FBtr0086024; name=CG7856-RA; dbxref=FlyBase:FBtr0086024,FlyBase_Annotation_IDs:CG7856-RA,REFSEQ:NM_136354; score=11; score_text=Strongly Supported; MD5=7ca9c4371b97aaf7046cf83fefb65eb8; length=2202; parent=FBgn0033056; release=r5.55; species=Dmel;
GGTAAAGTTGCCTTGGCGTCAGTTGGCAGTTTGGGAAAAGCCTACACACT
TAATATTTCGATAGATACACTTATTTCGCAATCGTAGAAGATACCACAAA
TCTCTCTTCCGTAAATTATAAGTATGTCCAAGAGGGTGAGCATCATGTTG
CCCGACGAAATACCTGCGGCTCCGTCAGGCAGCAGGAGGAACCCGATGCC......
So when I got the bam file, the column usually indicating chromosome localization is already filled with FBtr#:XXX-XXX (instead of chr1:XXX-XXX) which could be used as annotation.
I run the bam file in cufflinks to get the GTF file for each file and run cuffmerge to get a merged GTF file for files in the same set of experiment. Then I run cuffdiff with merged GTF file and bam files. I did NOT use any reference GTF files during those steps since the reference for mapping I used is already annotated as I described above. In cuffdiff output file isoform_exp.diff under the column of loci, the value also remained FBtr#:XXX-XXX. It means for each transcript, it is compared in different region. So should I consider if any of the region is differentially expressed, the transcript is differentially expressed? And is this going to affect validity of cuffdiff's calculation since this equals to do assembly first then quantification for differential expression compared to usually it does quantification first then assembly?
Thanks in advance.