Issues of ChIA-PET standards:

1. Are ChIA-PET data are going to be produced? How many of them will be?
2. What are the parameters about QC or others that are needed for ChIA-PET?

These two questions will need to be addressed by Dr Yijun Ruan for confirmation.

PLAC-seq:

Perform in-situ ligation like in-situ Hi-C to ensure library complexity, then use IP enrichment to decrease cost. Dataset has two components: short-range cis reads (<1kb), ChIP-seq signal; and long-range cis reads (>1kb), interaction detection. Normalization is needed because there may be systematic biases from IP enrichment: use local background to detect significant peaks, use methods from ChIA-PET to filter the non-contacting peaks, etc.

How to separate signal from backgrounds from antibody enrichment? The input is a ligated library of the whole genome (different from ChIA-PET, which is fragmented DNA) and IP is done via binding protein, therefore, the background would be much complex. Therefore, peak callers from ChIA-PET may not be entirely appropriate for PLAC-seq.

Up to a certain point, the data processing steps may be similar for PLAC-seq to Hi-C, therefore, some of the tools (like normalization) may still be applicable. Also, what kind of file formats would the new pipeline generate? Would it be one of the formats supported by current platforms? (Burak)

Known software, methods and formats will be considered first to process PLAC-seq data.

In ChIA-PET, comparison between two pools can be used for background for non-specific ligations, yet this is not available in PLAC-seq (Tim). PLAC-seq does not have the same level as ChIA-PET because ligation happens within the nucleus.

Things to do:

Dr. Ren’s group will put out an experimental guideline, data format and data processing procedures for PLAC-seq for the group to discuss about.

HiChIP from Howard Chang’s group, similar to PLAC-seq (sequencing technology is different but the biology is similar) may also be considered in this document.