Recommendations for creating TSS consensus clusters #1
Description
This may not be where you want to field this question -- please close this if so, and I'll e-mail.
I'm interested in creating a set of high confidence TSS for S288C from the STRIPE seq data.
From the paper, I see that you recommend a threshold of 3 counts per TSS, that a single 100-ng sample is used for purposes of making the figures in the main text, and in the supplement I see that the STRIPE seq replicates are very similar. From the code, I see the mapping of the STRIPE seq data in yeast_data/YPD_TSSs to the RNA amount/replicate.
Do you have a recommendation on how, which replicates, or whether to combine STRIPE-seq replicates to create a set of TSS annotations? Eg, just the 100-ng samples, take only those with >3 counts, and then create clusters/high confidence TSS regions where the replicate TSS are within 10bp?