ImmPort Ontology Conference

Where: Stanford University

When: September 4-5, 2013

Audience

• Day 1 is intended for all those engaged in information-driven immunology research who have an interest in the work of ImmPort and/or in ontology and data standardization
• Day 2 (by invitation only) is intended primarily for those interested in CyTOF and related issues of data management in immunological science.

Background resources

• An overview of ontologies proposed by ImmPort for use across the immunology research community is provided here

If you are interested in attending please contact Barry Smith as soon as possible.

Wednesday, September 4, 2013

Goals

• Support the work of ImmPort through advances in ontology development and use
• Provide examples of ontology content and of good practice use of ontologies which will help immunologists to rationalize their nomenclature and help them understand how ontologies are applied
• Work out with bench immunologists how nomenclature schemes can evolve to support enhanced discoverability and reusability through use of standards and ontologies
• Provide arguments and success stories that will help to achieve buy-in from bench immunologists as to the importance of standards and ontologies

Schedule

8:30 Registration and Continental Breakfast

9:00 Barry Smith (Buffalo): Overview of ImmPort Ontologies

9:45 Jeff Wiser (Northrop Grumman): Discussion on the Role of Ontologies in ImmPort

10:45 Break

11:00 Alex Diehl and Alan Ruttenberg (Buffalo): PRO, CL and CyTOF

12:00 Lunch

13:00 Holden Maecker (Stanford):Flow Cytometry Standardization and the Problem of Cell Typing

14:00 Melanie Courtot and Ryan Brinkman (Vancouver): The Cytometry-Ontology Framework

15:15 Break

15:30 Shai Shen-Orr (Tel Aviv): Ontology, NLP and the Semantic Enhancement of Immunology Research Literature

16:30 Lindsay Cowell (Southwestern Medical Center): Immunology Ontology and NLP

Thursday, September 5, 2013

We will begin by going through the steps of the ontological process involved in handling CyTOF data in order to address the following:

Major Questions for Discussion

1. Evaluation of the cell type definitions proposed in Maecker et al. (if possible do this prior to the meeting)

2. What should be the framework by which we can represent cell populations identified through given sorts of assays in such a way that we can later promote them to cell types recognized in the CL?

3. How do we determine what is really a new cell type rather than either a refinement of an existing cell type generated by additional markers, or (2) a transient activation state of some known cell type?

• Subtasks:

Ontological Background

a. explain the difference between continuant and occurrent

b. summarize how this difference is handled in other ontologies (especially GO)

c. list a set of properties that distinguish a cell type (continuant) from a transient state type (occurrent)

Metadata

What metadata need to be captured in order to enable the downstream determination that a cell population identified by some assay is in fact either (a) a bona fide cell type that should be included in CL, or (b) a bona fide cell state type?

• Proposed:

a. composition of the antigen panel

b. antibodies used to probe each antigen (expressed as ImmPort Antibody Registry ID)

c. the type of flow experiment: traditional, phosphoflow, CyTOF

d. unique experiment ID

e. species of the cells being probed (NCBI Taxonomy ID)

f. type of sample (whole blood, PBCs)

g. combination of markers that define a cell type according to the experimenter

h. clinical status of subjects (affected, unaffected; vaccinated/unvaccinated)

i. interventions (e.g., which arm of a trial the subject belongs to)

(At least in the short-run, it is anticipated that these data will be obtained from ImmPort's store of flow data.)

More ambitious questions for discussion if time allows:

3. Can we leverage CyTOF to develop a true step-by-step picture of hematopoiesis? This is a question for both ontology and the experimental approach.

4. What surface markers or internal proteins have reliable associations with biological processes, such that when we see a novel cell type or a variant of a known cell type we can predict the cell's function or (in other words the GO:Biological Processes it is capable of carrying out or participating in)? This question can obviously leverage existing GO annotations for particular proteins, some of which already have co-annotation with CL terms. But it can also lead to new terms for GO:Biological Processes and for CL cell types.

Schedule (Very Tentative)

8:30 Continental Breakfast

9:00 Nikesh Kotecha (Stanford / Cytobank): Software Challenges for Cytometry

9:30 Barry Smith (Buffalo / ImmPort): An Introduction to Ontology for CyTOF

What is a cell type definition? What is the difference between a cell type, a cell population, and a cell state?

10:00 An Introduction to Immunology for CyTOF

10:30 Immunology in the Gene Ontology (Alexander Diehl)

11:00 Alexander Diehl: The Cell Ontology (CL)

11:30 PRO

12:00 Lunch

13:00 Consequences for the Future of Immunology Science (See questions above)

16:00 Close

Participants

• Sanchita Bhattacharya (ImmPort / Stanford)
• Ryan Brinkman (Vancouver, BC)
• Atul Butte (ImmPort / Stanford)
• Quan Chen (NIH / NIAID)
• Lindsay Cowell (UT Southwestern, Dallas)
• Melanie Courtot (Vancouver, BC)
• Alexander Diehl (ImmPort / Buffalo)
• Sanda Harabagiu (UT Southwestern, Dallas)
• Nikesh Kotecha (Stanford)
• Suzanna Lewis (Berkeley)
• Holden Maecker (Stanford)
• Chris Mungall (Berkeley)
• Yannick Pouliot (ImmPort / Stanford)
• Alan Ruttenberg (ImmPort / Buffalo)
• Ravi Shankar (ImmPort / Stanford)
• Shai Shen-Orr (ImmPort / Technion Institute)
• Barry Smith (ImmPort / Buffalo)
• Nicole Vassilesky (OHSU, Oregon)
• Jeff Wiser (ImmPort / Northrop Grumman)
• Ashley Xia (NIH / NIAID)

Plus further participants from Stanford area.