4D Nucleome Network Wiki

AGENDA

Continue discussion with ChIA-PET data analysis

Continue discussion DCIC proposal regarding Hi-C visualization and process pipeline

ChIA-PET

Experimental has been available to everyone right now and last time discussion was visualization.

An electronic voting would be held to pass the proposal and the results are in favor. The ChIA-PET protocol will be forwarded to SC during the next SC meeting.

DCIC proposal regarding Hi-C visualization

Peter has a proposal that incorporate some of Erez’s advices.

1. DCIC’s own pipeline would ensure modularity
   1. In Erez’s proposal, the Juicer pipeline is already modular.
   2. Juicer pipeline is designed as an easy-to-use single package. While it is possible to edit the parameters in the code, it’s not designed to allow tweaking of the parameters. Codes are modular for sure but they are not designed to function independently. When something goes wrong in the middle, Juicer would require a start over from the beginning.
2. DCIC will reduce logistical challenges in incorporating the code into the DCIC framework.
   1. Erez proposed DCIC to fork Juicer and start from there, to reduce logistic burden on future development.
   2. Forking would essentially require code-rewriting and it would be more difficult to maintain a code base from someone else
3. The current Juicer does not conform to the data standards currently passed by OMICS group. While an update (v1.3) is coming and Erez proposed it to be conforming to the standard (and possibly being more modular), it is not available on the team Github yet.

The goal is to make DCIC’s job easier.

The current DCIC proposal does not include calling loops and domains.

• OMICS would discuss with the other working groups (for example, the Data Analysis WG) to seek advice on choosing the loop-/domain-caller, since there is no gold-standard for them yet.

An electronic voting would be held to pass the proposal and the results will be discussed during the next meeting.

It would be better for DCIC to pay attribution to the components in the pipeline by acknowledging citing the pioneering works under that component.

• DCIC’s pipeline would be open-source
• There can be a document on the 4DN data portal, including all works contributing the pipeline. This file can also be included in the open source code repository
• OMICS can also draft a protocol paper citing all the relevant a priori works.
• If all the codes available are going to be hosted on Github it would be easier to generate a contribution list.