Library normalization error. Also, peaks at non-genic sequences.

[juanb001]

Hello,

I have two questions.

1. I tried running CLAM with the --normalize-lib flag, but got the error in the attached picture. Is this an issue with the code?

2. It seems like CLAM calls peaks along the genes listed in the user-specified GFT file. Is there any way to call peaks at non-genic sites? The current method seems limited in its ability to call peaks at TEs that are outside of genes, since TEs are typically not included in GTF files, they would be skipped by CLAM. Or is there a way to call peaks at intergenic TEs?

[zj-zhang]

CLAM uses the gene annotations to perform a per-gene normalization. As long as TEs are within a gene (i.e. they are expressed as part/whole of an mRNA, not necessarily annotated in GTF), CLAM can call peaks on them.
For calling peaks on non-genic sites, if you have good prior knowledge about what you are looking at, it's possible to manipulate a custom GTF. Otherwise it is indeed undefined for CLAM to call peaks at non-genic TEs, because we don't have the proper background distributions to control for random spikes versus a true peak/binding signal.

[juanb001]

Awesome, thank you for the clarification.

Is library-normalization not recommended for RIP/CLIP-seq? And do you have an idea of what might have caused the problem in the picture above?