2014 12 10 Lifemapper team in Tucson

2014 12 10 Lifemapper team in Tucson

Organizers

Martha Narro, Ramona Walls, Nirav Mrechant

Date

December 10, 2014

Location

Keating 433, UA campus, Tucson

Purpose

Lifemapper is a platform for niche modeling. The iPlant and Lifemapper teams will discuss areas of common interest and identify the most fruitful ways to collaborate, leveraging what each project has developed.

Lifemapper Team

James Beach, Aimee Stewart, CJ Grady, Jeff Cavner

Lifemapper website

Schedule

9 or 10 AM

Start

12:00 PM

Lunch will be brought in

2:00 PM

Coffee Break

5:00 PM

End

Dinner

Lifemapper team will drive back to Phoenix and have dinner there.

Agenda Topics

  • Are there ways for Lifemapper and iPlant to work together to provide CI in environmental biology, perhaps linked through phylogenetic tree linked-data and into the geospatial/ecological domain, species niche or macroecological modeling?
  • Can iPlant leverage what Lifemapper has to enable users to perform species distribution modeling on HPC resources?
  • Is iPlant’s large tree viewer (phyloviewer in Discovery Environment) of interest/use to Lifemapper?
  • Discuss ideas for serving and visualizing large spatial datasets.
  • Is there data at iPlant that Lifemapper or its users would like to use?
  • What is each project's user community like?  Can we share?
  • Pipelines for niche modeling for agriculture
  • Joint funding opportunities
  • Workflows for assessing uncertainty in models
  • Submitting jobs to HPC: How Lifemapper does it. How iPlant does it (AGAVE API)

Meeting Notes

Notes in Google doc

Meeting Report

Nirav gave a technical overview of iPlant's infrastructure to enable the Lifemapper team to understand the CI and what they could leverage. The iPlant components Lifemapper is most interested in leveraging initially are authentication and the Data Store. This will alleviate some of CJ's pain points.

Lifemapper already has a well-regarded platform in place for niche modeling and a community of users, most of whom they collaborate actively with. Lifemapper has greater depth of expertise in the realm of niche modeling than iPlant. Ramona is the iPlant staff member with domain expertise. iPlant does not want to duplicate what Lifemapper already has built or compete with them, but rather to partner with them to provide any needed infrastructure support. Lifemapper is better positioned to engage directly with the ecological niche modeling community, with Ramona participating from the iPlant side.

When discussing the ecological/biodiversity organizations and community members each project works with, there was a great deal of overlap (NCEAS, Mark Schildhauer, Rob Guralnick, Walter Jetz, Pam Soltice). Lifemapper precomputes and archives species distribution models for GBIF. In the future they may precompute maps for data from BISON and iDigBio. iPlant has had greater engagement with the BIEN group, but Lifemapper is collaborating with a scientist at Harvard who is using BIEN occurrence data in a modeling project. iPlant has greater engagement with agricultural scientists interested in niche modeling support; Lifemapper team could provide expertise and a platform. 

Next Steps

  • CJ: Try uploading data into the iPlant Data Store, pulling it into Lifemapper for analysis and pushing results back to the Data Store.
  • All: Have a design discussion Dec. 19th at 11 AM CST, 10 AM MST. Martha will organize the call.
  • Lifemapper team: Will advise Nirav on how to put WebMap services on top of the Data Store. (current approach duplicates data, which is not desirable).