Tweet
Hello all.
I am using the Tuxedo workflow and trying to figure out the best parameters to use for my data. It's not easy, there are many options to decide whether to include or not include, and for a beginner it can also be difficult understanding them all. In the protocol paper they did not include many parameters at all, but I am sure some of them are quite useful.
My data is Illumina HiSeq 100bp paired end reads from a plant. I have a reference genome and annotation and have been assembling against the reference genome with the -gtf option in tophat. Mainly I want to focus on two conditions and test for differential expression between the two, but alternative splicing and novel splice sites, etc are also interesting.
Can you comment on some of the Cufflinks parameters and voice your opinion if you think they are appropriate for me to include?
So far I have been using Cufflinks mainly with these parameters:
--GTF-guide
--frag-bias-correct
--upper-quartile-norm
--multi-read-correct
--frag-bias-correct and --multi-read-correct seems like appropriate choices to include, but what about the other two? I still do not fully understand the RABT assembly and what it does and it seems like there are several normalization options, which one should I choose? Additionally I wonder if I should use the same parameters in every step in the Cufflinks workflow or only in the initial Cufflinks run? For example, many parameters like --frag-bias-correct you can use both in Cufflinks and in Cuffdiff. Cuffmerge has both a --ref-gtf and a --ref-sequence option. Is it a good idea to include both in Cuffmerge?
I am using the Tuxedo workflow and trying to figure out the best parameters to use for my data. It's not easy, there are many options to decide whether to include or not include, and for a beginner it can also be difficult understanding them all. In the protocol paper they did not include many parameters at all, but I am sure some of them are quite useful.
My data is Illumina HiSeq 100bp paired end reads from a plant. I have a reference genome and annotation and have been assembling against the reference genome with the -gtf option in tophat. Mainly I want to focus on two conditions and test for differential expression between the two, but alternative splicing and novel splice sites, etc are also interesting.
Can you comment on some of the Cufflinks parameters and voice your opinion if you think they are appropriate for me to include?
So far I have been using Cufflinks mainly with these parameters:
--GTF-guide
--frag-bias-correct
--upper-quartile-norm
--multi-read-correct
--frag-bias-correct and --multi-read-correct seems like appropriate choices to include, but what about the other two? I still do not fully understand the RABT assembly and what it does and it seems like there are several normalization options, which one should I choose? Additionally I wonder if I should use the same parameters in every step in the Cufflinks workflow or only in the initial Cufflinks run? For example, many parameters like --frag-bias-correct you can use both in Cufflinks and in Cuffdiff. Cuffmerge has both a --ref-gtf and a --ref-sequence option. Is it a good idea to include both in Cuffmerge?