4DN OMICS Standards Working Group

1/28/2019

Agenda

1. PLAC-seq/HiChIP discussion update (Miao Yu)
   1. Last November, we had a discussion/presentation focusing on IP efficiency, reads overlapping protein binding sites
   2. Since then, different batch of antibodies have been tested (K27ac, CTCF) for PLAC-seq.
   3. Briefly, for CTCF, three Abs were chosen. 1 from Cell Signaling and 2 from Santa Cruz. The one from Cell Signaling has the best yield of all the three. Santa Cruz antibodies appear to have lower binding affinity as well. Cell Signaling antibody is currently chosen
   4. The catalog number of the antibodies being chosen will be shared to the WG after the robust level of them has been assessed.
   5. Yijun Ruan presented ChIA-PET2 last December and it appears to be very similar to PLAC-seq and HiChIP
      1. PLAC-seq/HiChIP reduces noise level. But the antibody pull-down is observed to be more sensitive.
   6. The main difference between ChIA-PET1 and PLAC-seq/HiCHIP/ChIA-PET2 is in the order of ChIP and in-situ ligation. Therefore, the current discussion about antibodies for PLAC-seq and HiChIP is probably applicable to ChIA-PET2 as well.
   7. There are other differences in the analytic workflow, though. For example, ChIA-PET2 uses specific adapters so they needs to be seeked out during analysis.
   8. It is assumed that Yijun Ruan’s group will be using the updated protocol ChIA-PET2 within the ENCODE project. Erez will confirm this with Yijun. Once confirmed, OMICS will plan to add ChIA-PET2 with PLAC-Seq and HiChIP.
2. Single cell Hi-C sub-working group (Burak / Bill)
   1. There are two major types of single cell HiC experiments, sci-Hi-C and single cell HiC; and different single cell Hi-C protocols have different noise models that need different analysis filters.
   2. We agree on base data standards: fastq, bam, pairs
      1. sci-Hi-C uses one file per library (many cells) while other single cell HiCs have one file per library (one cell)
   3. The protocols for both methods continue to be improved, along with matching pipelines and QC criteria. Therefore no fixed protocol or QC criteria is recommended at this point.
   4. Single cell datasets are currently being generated by the JAWG and is part of the NOFIC-Allen Institute collaboration.
   5. DCIC has tested the sci-Hi-C pipeline and plans to implement it for open processing late 2019. (experiment types utilized by multiple labs are given priority for open processing implementation.)
3. Plan out the OMICS working group for 2019
   1. What techniques need to be discussed? And when?
      1. GAM -
      2. Methyl-HiC/Hi-Culfite
      3. Non-HiC related technologies such as GAM, SPRITE, both in the experimental site and the analysis side (SPRITE experiment protocol has already been approved and recommended).
         It is important to figure out how to integrate all these types of data. However the resource are limited at the Data Analysis WG side. This is within the purview of JAWG but it’s not clear who will be working on the integration.
      4. DCIC has tested both GAM and SPRITE pipelines provided by data generating labs and plan to implement them for open processing in 2019.
   2. How well are we meeting our goals?
      1. Shall we add Data analysis aspect to OMICS discussion
      2. Bill confirmed that this is a good idea.
      3. Next we can begin to schedule some discussion on the analytical side of methods.