---
tags: 4DN
---

# 2020-11-19 CMU+OME

Attending: Xiaopeng, Josh, Jason, Jian, Yang, Simon

## Intro
- Good meeting. Good people.

## Caterina on phase 1
- Jian: in phase 2 of DCIC (with David)
- JRS: also Niels G. (Jian: in dep't with Peter Park)

## Today
- Jian: type of data, prioritization
- how to visualize single-cell omic data (Hi-C), convert storm, do some comparisons, i.e. utilize capabilities & resources
- Jason: where to take notes, agendas. Everyone should see. Jian: FOMO
- But then: where will the datasets be for the consortium?
- Jian: phase 2 or just our project? Opinion: coordinate with DCIC but that's their job. But for our purpose, we can do something independently (and/or collaboratively) We should prioritize for our visualization, driven by discussions with Frank & Andy. Need to see certain things for their analysis based on their intuition.
- Data from the center. See table 3 from the proposal. Omic (manageable) & imaging. If in DCIC can acquire them periodically and update them in the interface.
- X. has some infrastructure for clicking on a track and triggering the OMERO viewer.
- JRS:
- Q1: where does the OMERO live and who's job is it to get the data in?
- Q2: ngff for sharing
- What's available?
- Budgeted some funds for CMU but not sustainable.
- Data storage should be the responsibility of the DCIC.
- Josh: public? non-public? s3? storage
- Jian: rely completely on the DCIC.
- JRS: Jason & Jian to talk to Niels, Peter, etc. "We have data. Are you ready? If yes, great. Otherwise we will need to figure something out" Need 2 years?
- Jian: Peter said that budget for phase 2 is smaller than phase 1. Made it clear that he doesn't want to make unrealistic commitments. (He let people go because of the budget) Overlap with HuBMAP? Zooming from DNA to nuclears to tisse to ....
- Jian: Priortize what will come out in this group. AICS wants to do their own solution. But make them interoperable.
- Josh: other projects with them, and quilt
- Jason: doing better than sending disks
- Jian: **Navigable Map** (interactive, multiple modalities) Global view of the nucleus, with interactions. Showing both chromatin nuclear bodies and something else. In a couple of years. ... other functions from AICS. Not just showing Hi-C & 3D genome structure. Putting them into a context of many other things to deliver a realistic model.
- JRS: labels/masks
- Targetting certain use cases
- Josh: n-dimensional spatial indexes (generic version of labels)
- JRS: we do the boring stuff
- Jian: did postdoc with Hauser (genome browser)
- pilot datasets demonstrating these capavilities
- paper where multiplexed storm looking at 3d & distance to nuclear bodies. Already published. Add new functions on top of that?
- Yang: only processed data. Downloaded. xyz positions of each loci.
- Jian: ask for the raw data. (Part of 4DN)
- JRS: "no one can handle the datasets"
- Jian: implement something first and then show them saying we want to connect.
- JRS: showing people what the benefit is. What's the scientific reason to do this? Finding ways to get more out of the data.
- JRS: there is an OMERO on the quad that she can import into.
- Perhaps queries of features on image.
- NB: Josh distances between objects
- X. http://genomewiki.ucsc.edu/index.php/Bin_indexing_system
- https://github.com/ome/omero-mapr/pull/61
- Jian: interval trees
- Josh: distance with some other constraint (Jian: interesting queries)
- What should we deliver first? (Use cases)
- e.g. no one cares if you show one image (Will/Jason)
- Next meeting in December 16