# Wrapup

## Team: RPhenoSCATE

### Goals

[Github repo](https://github.com/uyedaj/rphenoscate)

A matrix from ontotrace that is annotated with UBERON

Can you estimate if the trait was evident in the ancestor?

Phylogeny trees w/ traits on them, shows better understanding of whether it is possible for a trait to be inherited

Shiny app to generate these from list of traits that generate phylogeny that have the trait.

Shiny apps can be a good way to do education.

See difference between ontologically-informed ancestral state reconstruction vs ontologically-uninformed ancestral state reconstruction.

Based on paper Sergei has in [biorxiv](https://www.biorxiv.org/content/early/2017/09/19/188672)

### Questions

* Are there web services Phenoscape should provide to make these? - given a list of characters, make a matrix
* R package rontology - requires reloading the ontology
* Ancestors may be more than 2 types

## Team: Map ott to vto

Looked an catfishes taxonomy 4763 taxonomic concepts, 3265 have a 1:1 match between the two, 29.9 % have no match

Displayed on an OpenTree tree - a few cases with 3-4 matches in VTO

Examples of 1 OTT to multiple VTO: [*Ompok bimaculatus*](https://tree.opentreeoflife.org/taxonomy/browse?name=Ompok+bimaculatus) synonyms but valid, more investigation needed

Wanted to look at all VTO too, used NCBI ids in VTO which used as a common identifier - 40% map succesfully, 48% map through TNRS with exact name, 3% synonyms, 9% don't match

Going from Opentree to VTO - many cases where OT ids match to many VTOs
- One open tree ID maps to 10 VTO names!
- "Amyda radula"
- Extinct taxa in GBIF is likely where these names come from

Wrote to a triplestore, add OT IDs as a property to the taxa in phenoscape

Mapping is in text file [on Github](https://github.com/phenoscape/phylopheno)

### Question

* Does this only happen when OT is a species?
* How to convert to RDF - there is a script in the repo to_rdf_

## Team: iDigBio, specimens and Phenoscape
* Github: https://github.com/phenoscape/KB-DataFest-2017-linking-data
* ~3,000 specimens in Phenoscape matched with things on iDigBio using specimen IDs
* So they used taxon names instead.
* Wanted to make heatmaps to show how characters varied geographically
* And also wanted to show how they varied on phylogenies
* Components:
* Inputs:
* Select taxa: VTO IDs from Phenoscape, which may be species or higher taxon
* Select character that Phenoscape has for that taxa
* Heatmap:
* Summarized by genus
* Displays number of specimens for each taxon in iDigBio
* Phylogeny:
* Not hooked up yet
* Currently based on two character matrices, but long term should connect to the Phenoscape APIs
* Only ~110 tips out of 6,500 names from Phenoscape match in opentree currently

## Team: Phenoscape For Plants
- Laurel, Annika, Lorena, Hilmar, Jim
Four Components:
- Plant Phenoscape platform
- Phenex for plants
- Arabidopsis model plant data for KnowledgeBase- from Oellich paper (https://plantmethods.biomedcentral.com/articles/10.1186/s13007-015-0053-y)
- Data matrices generated from plant monographs through Hong's ETC tool converted to Nexml in Mesquite

Products:
* Configured Phenex with custom set of ontologies for plant annotations
- Using NCBI taxon for just plants
- Plant Ontology
- Plant experimental conditions ontology
- PATO, RO

* Arabidopsis model plant data, had mutations parsed into EQ statements already, configured these to match zfin format
* Set up GitHub repo for data files on Planteome GH (https://github.com/Planteome/plant-phenoscape-data)

Have everything in position to make plant phenoscape set up

Also working on reproducible build process for stand alone/additional versions of Phenoscape - have Owlry component of Phenoscape dockerized on dockerhub, can pull and run the container on your local machine, can query owlry once running via web service in DL query

Challenges:
* Still working on orchestrating multiple containers

* Making the matrix in the right format is hard - still working on 400 taxa to extract information - from publications and pipeline for parsing monographs into matricies, not all terms in an ontology

## Team: Phenopackets

What are the traits that occur in low light environments? traits and environments

RDF description of environments, based on phenotypes in wasps, quering data in the packages is difficult - the queries are long

Tried to design a phenopacket schema - moderatly successful - but think triples in a store are easier to query. Packets may make a good transfer mechanism, not a store

Demo of the packet schema - light, temp, pressure of haddle zone

Reference implimation didn't allow validating schemas because they were hard coded. Wrote a lot of transformations themselves

Have wireframes of how querying by environment in phenoscape might look.

Design pattern for generating ontology terms - YAML - end user could use Noctua - convert this into OWL to be injested

Packet design isn't easy to work with

### Questions:

* How do you make th emock up? Illistrator

## Team: NeXML to JSON-LD

Improve interoperability

Implemented python convertor, ruby convertor almost working

Developed notes on NexLD version 2 for better conversion

Low level Ruby client for Phenoscape API

Fishtank: Tool to produce svg. uses spatial coordinates on UBERON parts, and render a visualization by part, by taxon by quality.
Uses the NeXML and the api and the Ruby.

### Questions:

## Team: Noctua for phenotype annotation

Some challenges making it publicly available.

But it is not up at:
noctua.phenoscape.org

Request account from Jim Balhoff if you would like to try it out.

Created a graph to link together specimens from differnet environments with differnet traits.

Hope to have a query to determine which traits are linked to different environments.
Summary: Everything is possible with Jim.

# Sum up:
Everything that we did here should be made available: GitHub repo, file share, etc.

Look at subgroup page in the wiki, and link out to your documentation.

Document!

Hilmar will enforce

### Things we liked:
everyone active and participating
morning standups
location space, venue
interplay between groupd
OpenSpace: the way of finding the working groups/breakout process
People got plugged in rathe than being lost
Focus on a single API reuslts in synergy
Jim's API bootcamp
Jim.
working with people we've known but don't get to hang out with
Eveyone was open minded and forgiving of API challenges
Exposure to details of the porject, absorption of others' ideas
Breadth of people. Not a bubble!
Breadth within groups as well as across whole workshop
Schedule worked, working + socializing
Good number of people. Felt right.
Rice crispie treats + the fruit + the chocolate
Other rooms available for breakout sessions
Tutorials were really helpful for beginners

### Things we liked less/should change:
Lunch in on the last day
Wrap up felt time crunched
Had trouble prepping in advance - website was unintuitive, and web APIs were challenging.
Was hard to know if properly prepared or not. A pre workshop activity, or primer might be helpful, and open up more people to pitch
Potentially identify first time participants and reach out to them?
Documentation of the project would help clarify