skipper applicability

[mfurla]

Dear developers,

Thank you for this great tool.

I would like to try skipper to detect binding sites in a HITS-CLIP (not eCLIP) experiment for which I have IP and IgG samples in two biological replicates, sequenced with unstranded libraries.

Would it be suitable? If so, could you recommend the appropriate settings to use?

Best regards,

Mattia

[augustboyle]

Hello Mattia,

IgG controls usually capture a small number of events that are low complexity, very highly expressed, or technically biased for some other reason. The ones I have looked at don’t resemble a matched input. I have also attempted to use published RNA-seq from another group but that also differs greatly, either because the protocol or exact cell sample is different.

If the IgG sample has very sparse, diffuse reads, then background is low, and Piranha called peaks should perform quite well. The more signal the IgG sample has with apparent peaks throughout the genome, the less reliable the Piranha peaks will be. In that case using the IgG sample as a covariate for Piranha and using them as inputs for Skipper might be worth trying, although I would carefully inspect the output. Skipper might fail to fit overdispersion parameters and either crash or yield no hit windows.

[mfurla]

Hi August,

Thank you for your quick and informative response!

From a manual inspection of the IP and IgG coverage tracks, they appear very similar in terms of signal, so I think I will follow your suggestions and try both Piranha with a covariate and Skipper.

Regarding Skipper, do you foresee any issues related to the CLIP protocol or the sequencing strategy used? These experiments are HITS-CLIP (not eCLIP i.e., not enriched in truncated reads) and were sequenced with unstranded libraries.

Thank you again for your availability and support.

Best,

Mattia