Felix Krueger

My guess is that you may want set it to "Maximum" difference in quantitated value instead of average...

Hi Jonathan, I would say this is the first time in well over a decade that I see anyone using FastA files for alignment. I am wondering how you got...

Your command contains the flag `-f` which instructs Bismark to expect FastA files. Just drop it, and all will be fine.

That's all fine, you just have a few reads that align to the very end of the mitochondrium. Just keep going!

Hi @franrodalg I have tried to make a slide explaining this [here (slide 27)](https://www.bioinformatics.babraham.ac.uk/training/Methylation_Course/BS-Seq%20theory%20and%20QC%20lecture.pdf). Spelled out it would translate to: - if a sequence contains the the entire 13bp start...

Hi Fran, apologies for the late reply, but Easter egg hunts got in the way... I don't think I quite agree with the statement that "fragment size is [largely] driven...

I personally would not filter out anything from the start (i.e. at the trimming step). If you wanted to test the effect of different fragment lengths you could always do...

I didn't think you were, that would be really bad :) But yes, methylation data is extremely bi-modal, with >> 90% of data either being highly methylated (80-100%) or not...

Agreed, reporting single average value for something like methylation data is - interesting. Histograms or violin/beanplots are indeed much more informative, see e.g. here [slide 29]: https://www.bioinformatics.babraham.ac.uk/training/Methylation_Course/Visualising%20and%20Exploring%20Lecture.pdf

Working with common positions that were covered well in all samples would be ideal, and get you around looking at coverage issues such as in slide 22 in the above...