Andy Deans

Undergraduate opportunities in phenotype research

 Data Curation, Informatics, Jobs, Morphology, taxonomy  No Responses »
Mar 012013
 

 

We’re looking for a few good interns! Photo by Gilles San Martin. (CC BY-SA 2.0)

The Frost Entomological Museum at Penn State seeks undergraduate summer (2013) interns to assist with projects related to insect phenotype data, especially in the context of systematics and evolution. Interns will be exposed to a broad array of biodiversity informatics tools, including ontologies, and will learn aspects of specimen collection, handling, and curation. Applications are due March 20, 2013. More information is available at http://bit.ly/FrostInterns

Arthropod working group update – fall 2012

 Anatomy Ontology, HAO, Morphology, OWL  No Responses »
Nov 202012
 

Asilomar State Beach, Monterey, California, USA. The beach served as our chairs, and the rocks were our whiteboard.

We had a short but highly successful meeting a couple weeks ago, at the Asilomar State Beach conference center in beautiful Monterey, California. Our working group, the AAO, met in parallel with the vertebrate anatomy (with a twist of anthropology), plant anatomy, sponge morphology, and informatics working groups, which allowed for a few useful group exchanges and productive collaboration. (And the setting was difficult to beat! See above.) The AAO set several ambitious goals for this meeting, and we made substantial progress towards a first draft of a common Arthropod Anatomy Ontology. Arthropod group accomplishments from this meeting include:

  1. We moved closer to a general common understanding of how we need to build a broadly useful common arthropod anatomy ontology. It can be difficult to reach consensus and understand through Skype and email. The opportunity for a face-to-face was invaluable for our group as we move forward in this effort, especially since our group includes experts from multiple disciplines (primarily genomics, comparative morphology, and systematics) and is prone to unintentionally talking past one another.
  2. We moved closer to having a shared environment to view, search, and safely edit a common ontology with integrated OWL reasoning. We had been developing two ontologies in parallel, one in Protégé and one in mx, and one goal of this working group meeting was to reconcile those ontologies and move to a single mechanism for ontology development. We’ll be moving forward with an extended version of mx (developed over the next several months, hopefully) that includes a continuous integration and testing environment and which will automate error checking and feed back inferred classifications to mx so that they are visible for viewing. Ideally this environment will be extended to allow queries.
  3. We made some headway towards recruiting more experts and raising awareness of our effort. We have a short list of other experts to target for face-to-face meetings and ideas for how to bring this project to potential consumers through meeting talks and/or symposia. Stay tuned for more details!
  4. First draft definitions for general classes of arthropod external anatomy and a first attempt at OWL formalization for them (our biggest accomplishment). We spent a lot of time on cuticular classes (e.g., sclerite, conjunctiva) and articulations, including the various parts of articulations—structures we’d call condyles, fossae, etc. Some classes were relatively uncontroversial, for example articular surface: A sclerite surface that makes movable direct contact with another articular surface, with the properties is_a sclerite surface and contacts articular surface. Other classes were much more challenging, and our discussion often veered into the term/class danger zone. ‘Joint’ was especially troublesome. See for example the ‘wing joint’ of pterygote insects, which is a highly complex anatomical entity (image below). Another controversial (reactive? radioactive?) class was ‘area’, from the Hymenoptera Anatomy Ontology (HAO): An anatomical structure that is delimited by material or immaterial anatomical entities. Is this class useful for Arthropoda (or even Hymenoptera)?

Proximal area of a sawfly forewing, where the wing meets the thorax. This is a highly complex region of the body, chock-o-block with sclerites, conjunctiva, muscles, etc. Is there a way to define ‘joint’ so that it satisfies this complex? Should it be called ‘joint’?

As mentioned above we currently have two draft arthropod anatomy ontologies—one built largely by David Osumi-Sutherland (FlyBase, Cambridge University) in OWL and the other built largely by István Mikó (HAO, Penn State) in mx as a generalized version of the HAO—which overlap substantially, especially in general classes for external anatomy. David has been reconciling these two versions in one OBO-format file, with the aim that this will be used to seed the common version in mx (see small test ontologies on GitHub if you’re interested).

We’re considering a couple slight changes in our approach going forward: 1) moving to multiple meetings a month, rather than just one, and 2) establishing semi-independent subgroups that share anatomical interests, which could lead to more substantial growth/refinement of relevant areas (e.g., the nervous system or circulatory system). Email me, Andy Deans (adeans@psu.edu), if you are interested in the latest developments or to find out how to contribute. Thanks to all who were part of this working group: Frank Friedrich, István Mikó, David Osumi-Sutherland, Aaron Smith, Shaun Winterton, and Matt Yoder!

Arthropod Working Group update

 Anatomy Ontology, HAO  No Responses »
Aug 202012
 

One of the many goals of the Phenotype RCN is to generate a common anatomy ontology for Arthropoda, the phylum that includes “segmented” organisms (more on that in another post), with “jointed” appendages (again, more later) and a rigid exoskeleton (and more on this later). We’re talking about spiders, scorpions, centipedes, crabs, shrimp, insects, etc., a lineage that includes numerous model organisms (e.g., Drosophila melanogaster) and which accounts for more than a million known species – more than 50% of all organisms! Given the diversity of body forms exhibited by these critters, not to mention the huge array of researcher communities, one can imagine the disparity of concepts and terms for their body parts.

Several arthropod-relevant, taxon-specific anatomy ontologies have been developed in the last 10 years: one each for Drosophila, mosquitoes (Diptera: Culicidae), ticks (Acari), spiders (Araneae), and Hymenoptera (Insecta). Each was developed more or less independently, with no coordination of common, higher-level classes. Yet each community of researchers—the arthropod-relevant anatomy ontology users—stands to benefit from a set of common set of shared, higher-level classes. Imagine connecting one’s observation of millipede antenna phenotype to the vast body of Drosophila mutant phenotype data and also to a scorpion species description, through a common, higher-level ‘appendage’ class. Several domain experts are now working on new arthropod-relevant anatomy ontologies, especially for Neuroptera (lacewings), Coleoptera (beetles), and, more specifically, Tribolium. Should they develop  their own ontologies, or should  their efforts benefit from the experience, standards, technology, workflow, philosophy, classes, etc. of existing efforts? Here’s where the RCN comes in.

Psychopsis insolens by Shaun Winterton
How can we connect this ant lion’s phenotypes (e.g., the fuzzy wings) to knowledge derived from model organisms (e.g., Drosophila and Tribolium)? Perhaps the AAO can facilitate these connections. Photo of Psychopsis insolens (Neuroptera: Myrmeleontidae) by Shaun Winterton.

We’ve formed an arthropod working group (AWG), whose mandate is to identify and define these common classes across major arthropod lineages. We met face-to-face once, last summer at the Google offices in Boulder, CO (notes here). We also met briefly during the last summit meeting (blog post here), where we got a bit more organized about how this effort should look. And finally, true to our word, the AWG has been and continues to engage in regular skype chats – 3rd Friday of every month at 5:00pm GMT – and they seem to be working quite well.

A few of our accomplishements so far:

  1. We have a wiki, which includes a list of relevant projects and current AAO contributors: http://aao.speciesfile.org/
  2. We have an email list: arthropod anatomy discussion
  3. And a term / concept request tracker: arthropod AO
  4. We also tackled several higher level concepts, including sclerite, appendage, appendage segment, segment, tagma, and glands, and identified relatively uncontroversial areas of the ontology could grow without too many arguments.

The successes and failures of our first few group discussions have helped us refine our workflow, and we now have a plan for partitioning the effort more effectively. We anticipate substantial refinement of the current classes at our October working group meeting, as well as considerable growth in the depth of our ontology. We’ll report the results here! In the meantime please peruse our emerging resources and contact Andy Deans (adeans @ gmail . com) if you’d like to get more involved.

 Posted by on August 20, 2012 at 2:49 pm

Arthropod Working Group at the Phenotype RCN

 Anatomy Ontology  No Responses »
Mar 052012
 

External male (left) and female (right) genitalia of ceraphronoid wasps, as visualized using laser confocal microscopy. Which parts of these anatomical complexes are true 'appendages' and why? How do we define an arthropod 'segment'?

It’s been just over a week since our grand Phenotype RCN summit meeting at NESCent, and now that the dust has settled it’s a good time to report on some progress and future directions of the Arthropod Working Group (AWG). The AWG is one of four somewhat loose-knit working groups, organized under the Phenotype RCN umbrella, and our primary goal is to develop a common anatomy ontology for Arthropoda (CAAO, for now). This summit meeting was the third opportunity for us arthroöntologists (pretty sure that wasn’t a real word, until now) to meet since the RCN began last year, and we capitalized on the positive vibe permeating through the greater meeting to discuss five main questions:

1) What resources do we have already to move this effort forward in an organized, robust way? We itemized the various communication and ontology development tools we have at our disposal: Arthropod Anatomy email list (not used much yet, but could blossom), an arthropod anatomy ontology Google Code site with issue tracker, ontology development tools (primarily Protégé and mx), and Skype (of course!)

2) Given these resources, how do we envision the workflow unfolding? We’ve been working in a somewhat piecemeal fashion for the last year, which made progress a bit … slow. Given the varied informatics expertise of our group we still need to discuss mechanisms that will allow everyone to contribute in a relatively flexible way – e.g., some people could checkout the ontology and refine it using Protégé, whereas others will want simply to send requests via email. This discussion is ongoing.

3) Ontology development should be guided by real use cases. What are ours? Our working group (n=17 so far) is populated by systematists, comparative morphologists, neuro-anatomists, and evolutionary developmental biologists. And our study organisms ranged from established models (Drosophila) to emerging models (Parhyale, Gryllus) to fossils. Accordingly, our use cases were quite diverse: connecting behavior to anatomy (For any given behavior, what structures are implicated in clades X and Y?), discovering homology across Arthropoda (e.g., topologically same structure but named differently in different clades), understanding the context of internal arthropod anatomy (in which clades are the ovaries dorsally located anterior to segment X?), standardizing descriptions of comparative anatomy results (making descriptions explicit and logical), query for data about homologous or similar anatomical structures in related species (e.g., gene expression, phenotype, electrophysiology data), make data collected in one domain available/queryable to all other relevant domains (e.g., from descriptive taxonomy to genome-phenome researchers).

4) Given these use cases, how do we prioritize ontology development or otherwise partition its development? This discussion is ongoing, but we tossed around options for recruiting more experts to contribute for certain high priority anatomical systems. E.g., if we see immediate payoff for implementing the CAAO in the context of ovarian phenotypes we could organize a push to recruit more arthropod ovary experts for a working group.

5) What are our action items? We definitely need to organize regular virtual meetings, either as Skype chats/calls or regular conference calls. We’re scheduling the first chat soon and will post here when we finalize the date/time!

So, that’s the quick summary of our working group meeting. Expect more news soon, right here on this blog. In the meantime, if you crave a more detailed report of the AWG activities so far, and you seek to get more involved, don’t hesitate to contact Andy Deans (adeans @ gmail . com) – especially if you think you know the answers to the questions highlighted in the image caption above!

 Posted by on March 5, 2012 at 8:30 pm

RCN working group meeting – Boulder, CO

 Uncategorized  Comments Off
Aug 272011
 
Last June the Phenotype RCN organized a set of concurrent working group meetings—focused on arthropod anatomy, vertebrate anatomy, plant phenotypes, and the informatics associated with phenotype data and ontologies—that were very graciously hosted by by Google's office in Boulder, CO. The meeting agenda and subsequent notes are already posted on the Phenotype RCN website, though I noticed today that the arthropod anatomy working group summary never got posted!

The longer version is available as a Google doc. Here's the short version: We are working towards developing two higher-level anatomy ontologies, one for classes in common across Arthropoda and another for Insecta. We've identified a series of potential classes, proposed candidate definitions for notoriously difficult ones (what's a 'segment'?!), identified some tools we can use to facilitate communication (an email list, Dropbox, Google docs, Git), development (mx, Protégé) and visualization (C-Map, Vue), and identified several highly motivated people to collaborate with.

I'll also use this opportunity to thank our host at Google's Boulder office, John Day-Richter, for hosting what ended up being an inspirational, productive meeting! Our next group meeting is at the National Evolutionary Synthesis Center in Durham, NC. Watch for more details!