Tweet
Hi Everyone,
I have simple small question in regards to Cuffmerge2 and Cuffdiff2. I had two different experimental sets from NCBI Geo.
Let's call them Experimental Set A and Experimental Set B.
The Experimental Set A was phred33 (Sanger/Illumina 1.8+) and Experimental Set B was phred64 (Illumina .18).
I ran Tophat 2 and Cufflinks 2.0.8. Everything has worked well up to know. I did not use Cufflinks 2.1.1 because of the high and low confidence interval bug in the application.
Now I was wondering should I run them with Cuffmerge 2.0.2 or Cuffmerge 2.1.1 on them and likewise for Cuffdiff2.1.1 or Cuffdiff2.0.2? Does it really matter which one that I use?
On top of this, should merge both experimental Set A and Experimental Set B's Cufflinks files together? Does it really matter what their original phred score was they have gone through Tophat2 and Cufflinks2?
What are the recommended parameters for Cuffdiff2 for comparing gene expression between two organs in the same species.
Is 10 a good value for Minimum per-locus counts for significance testing?
Should I perform a multi-hit correction? I did this for Cufflinks in the past, but I never went further to Cuffmerge2/Cuffdiff2/Cummerbund.
Should I normalize the data based on the known hits to transcript or should I go with total hits for the normalization process. What's the difference?
What value should I place for the recommended False discovery rate value and for Jensen-Shannon replicates? I was thinking of stating 0.5.
Is 0.1 a good value to place in for Cuffmerge2's Discard Isoforms with abundance below this (0-1)?
Sorry for the so many questions. Its my first time doing Cuffmerge2 and Cuffdiff2.
Then I hope to use Cummerbund through R to see the difference of gene expression for specific genes of interest for us.
Thank you in advance.
I have simple small question in regards to Cuffmerge2 and Cuffdiff2. I had two different experimental sets from NCBI Geo.
Let's call them Experimental Set A and Experimental Set B.
The Experimental Set A was phred33 (Sanger/Illumina 1.8+) and Experimental Set B was phred64 (Illumina .18).
I ran Tophat 2 and Cufflinks 2.0.8. Everything has worked well up to know. I did not use Cufflinks 2.1.1 because of the high and low confidence interval bug in the application.
Now I was wondering should I run them with Cuffmerge 2.0.2 or Cuffmerge 2.1.1 on them and likewise for Cuffdiff2.1.1 or Cuffdiff2.0.2? Does it really matter which one that I use?
On top of this, should merge both experimental Set A and Experimental Set B's Cufflinks files together? Does it really matter what their original phred score was they have gone through Tophat2 and Cufflinks2?
What are the recommended parameters for Cuffdiff2 for comparing gene expression between two organs in the same species.
Is 10 a good value for Minimum per-locus counts for significance testing?
Should I perform a multi-hit correction? I did this for Cufflinks in the past, but I never went further to Cuffmerge2/Cuffdiff2/Cummerbund.
Should I normalize the data based on the known hits to transcript or should I go with total hits for the normalization process. What's the difference?
What value should I place for the recommended False discovery rate value and for Jensen-Shannon replicates? I was thinking of stating 0.5.
Is 0.1 a good value to place in for Cuffmerge2's Discard Isoforms with abundance below this (0-1)?
Sorry for the so many questions. Its my first time doing Cuffmerge2 and Cuffdiff2.
Then I hope to use Cummerbund through R to see the difference of gene expression for specific genes of interest for us.
Thank you in advance.