ICBO 2014 in Review

We wrapped up ICBO 2014 late yesterday afternoon with an "unconference" session, whereby attendees guided the topics and discussion of the final hour of the program prior to the closing session.

A couple notes:
1. Fred Trotter requests a "10 commandments" of ontology development.
2. Jim Overton says our tools for ontology development and curation are in terrible shape.  He received a round of applause, and it wasn't the applause that came at the end of his two minute spark talk.  It was instead widespread recognition of the problem.

Congratulations to Dr. Werner Ceusters, who won the best paper award for his paper and presentation on Pain Assessment Terminology in the NCBO BioPortal: Evaluation and Recommendations.  The award is sponsored by the National Center for Biomedical Ontology, and Dr. Ceusters will receive a $500 prize for the award.

Thank you once again to our sponsors.  In addition to NCBO, we thank NASA, Apelon, and DiBS (Data is Beautiful Solutions).

Thank you once again to our supporters.  We could not have had such a well-oiled, efficient, smoothly-running conference without Health 2.0 Houston.   We also received support from Houston Technology Center, and the School of Biomedical Informatics at UT Houston Health Science Center.

The 5th International Conference on Biomedical Ontology kicks off tomorrow

I'm here in Houston with my co-organizer, Sivaram Arabandi, sitting outside on a beautiful day as we put the finishing touch on what promises to be an outstanding program for the 5th International Conference on Biomedical Ontology (ICBO14).

ICBO14 kicks off Monday, Oct 6th with workshops and tutorials.  We have four--yes four--workshops that deal with the ontology of drugs and drug interactions.  There are also workshops on the ontology of biobanking, the OBO Foundry, and definitions in ontologies.  We also have tutorials on the OBO Foundry and mapping relational databases to the Resource Description Framework.

Our keynote speakers are highly distinguished.  Dr. Warren Kibbe--currently the the director of the Center for Biomedical Informatics and Information Technology (CBIIT) at the National Cancer Institute (NCI)--will give a dinner keynote on October 8th entitled Supporting Precision Oncology using Ontologies.

Also giving a keynote address is John Wilbanks--Chief Commons Officer of Sage Bionetworks, who will talk about Ontologies and Open Collaboration.  John has also worked at Harvard’s Berkman Center for Internet & Society, the World Wide Web Consortium, the US House of Representatives, and Creative Commons. John is a past affiliate of MIT’s Project on Mathematics and Computation and also started a bioinformatics company called Incellico, which is now part of Selventa.

We have a great scientific program with full-length and early-career papers, posters, software demonstrations, invited updates, and panels.  We are especially looking forward to a panel on ontology and genomics with Christian Stoeckert, Bjorn Peters, and Jim Zheng of the Center for Computational Biomedicine at the School of Biomedical Informatics here in Houston.

If you are not already registered or planning to come, you're going to miss out on a fantastic program!

Age is not a quality

No matter how many times lazy ontologists represent age as a quality of a human, organism, or other material (or immaterial I suppose) entities, it just isn't true.  Age is not a quality.

An age is a measurement of elapsed time since a particular entity came into being.  Typically, it is the measurement of elapsed time in the frame of reference of the entity itself.  For example, muons have a lifetime of 2.2 microseconds as measured from the frame of reference of the muon itself.  However, as measured from the frame of reference of the surface of the Earth, the muons entering the atmosphere have a mean lifetime of approximately 18.5 microseconds.

Therefore, age should be a type of measurement in ontologies.

We also note that age is always with respect to the time the measurement is made.  For example, if John Doe was born on Jan 1, 1970 UTC, then if we measure his age on Jan 2, 1970 UTC, we will obtain that John is 24 hours old.  But if we measure his age on Jan 1, 1971, we will obtain that he is 365*24 =8,760 hours old (1970 not being a leap year).

So an age measurement should always be accompanied by an appropriate measurement date/time.

Plant Ontology and Ontology of Biomedical Investigations Admitted to OBO Foundry

The OBO Foundry has announced that it has admitted two additional ontologies to its membership.  Specifically, the Ontology of Biomedical Investigations (OBI) and the Plant Ontology (PO) met the standards of the OBO Foundry Editorial Working Group.  This move brings the number of Foundry ontologies to ten:

1. chemical entities of biological interest
2. biological process (gene ontology)
3. cellular component (gene ontology)
4. molecular function (gene ontology)
5. protein ontology 
6. phenotypic quality
7. xenopus anatomy and development
8. zebrafish anatomy and development
9. plant ontology
10. ontology of biomedical investigations

Per the announcement, ...OBI and PO have recently been reviewed by the OBO Foundry Editorial

Working Group, and reviewers found that they performed well when measured against accepted OBO Foundry Principles... Consequently, the Editorial WG recommended that PO and OBI be given full

member status as OBO Foundry Ontologies. This decision was ratified by the OBO Foundry Operations Committee.

Progress towards high-quality, peer-reviewed ontologies deemed acceptable for use in scientific and other endeavors thereby has taken a strong step forward.

Glossary designers cannot define components of glossaries

The International Organization for Standardization (ISO) is working on standard TS 17439 Health Informatics Common Glossary Metadata Repository and Maintenance Process, which aims to define the key components of glossaries in health informatics, so that all such glossaries will be interoperable.

However, the glossary people are having a hard time defining the key components of glossaries.

For example, the definition of 'term' is: The word or group of words being defined in the glossary.

This definition implies that terms do not exist unless they are in a glossary, which is clearly untrue.  Terms existed long before people put them in glossaries.

The definition of 'term ID' is:  A computer generated unique identifier for this instance of the term.

Why is it a necessary condition for term IDs that they be computer generated?  What if a human manually creates an identifier for a term?

Furthermore, what does it mean to have different "instances" of a term?  It appears that it means the appearance of a term in a particular health standards document.  So if the same term appears in two different standards with different definitions in those standards, it gets two identifiers.  But then how does one know that it's the same term?  Uniqueness of the string of characters that make it up?  But then how will one compare the definitions of 'arm pain' and 'pain in the arm'?

The definition of 'context' is: Specialisation context

That is like defining 'automobile' as 'blue automobile'.

Hopefully the glossary makers can get their own definitions right before this proposed standard is approved.

Cardiology Society (non) Defines Key Entities for Cardiovascular Care

The American College of Cardiology (ACC) and the American Heart Association (AHA) are to be lauded for their goal of creating standard definitions of key terms for use in collecting scientific data.

However, their execution is--quite frankly--abysmal.

Here are some definitions they generated:

Sex	Indicate the patient's sex at birth as either male or female. Choose 1 of the following: Male Female
Date of birth	Indicate the patient's date of birth.

These are NOT definitions at all.  We have been misled.  These are instructions, however obvious.  What did the ACC and AHA think someone might do when confronted with a data entry form asking for the patient's sex and date of birth?

For real definitions of sex and date of birth, and a standard way of recording them that requires no special data machinery specific to demographics, see a paper by Hogan et al. from the 2011 International Conference on Biomedical Ontology.  These definitions have been captured in the Demographics Application Ontology, available here.

But since the real goal is to capture data elements specific to the field of cardiology, perhaps it is too unfair to criticize their efforts on demographics.   Cardiologists, after all, are more concerned with the heart than the rest of the person.

Alas, although they do move from instructions to definitions, their definitions are poor.  They are circular, meaning that they use the words in the term to define the term:

Prior angina	History of angina before the current admission. “Angina” refers to evidence or knowledge of symptoms before this acute event described as chest pain or pressure, jaw pain, arm pain, or other equivalent discomfort suggestive of cardiac ischemia. Indicate if angina existed >2 wk before admission and/or within 2 wk before admission.
Average number of episodes of angina in the prior week	Average number of distinct episodes of anginal pain that occurred in the last week before hospital admission or this visit

Furthermore, what if the patient is not being admitted to the hospital?  According to this definition, a clinician should not record "prior angina" if the patient had a history of angina prior to a clinic or emergency department visit.

The definition of "angina", alas, although promising, is ruined by the prefix "evidence or knowledge of".   If there is no evidence or knowledge of Mr. Smith's angina (perhaps he has no memory of it due to some mental disorder such as dementia), it still exists.  It also conflicts with the subsequent instruction to record the existence (not knowledge or evidence) of angina.

Furthermore, it really defines two data elements, despite being listed as one.  We are told to record existence (not knowledge or evidence) of angina before 2 weeks prior to admission (data element #1) and the existence (not knowledge or evidence) of angina within 2 weeks prior to admission (data element #2).

Thus the data element "prior angina" is also ambiguous.

The circularity of the definition of "average number of episodes of angina in the prior week" is self evident.  Note that now we are told that by "admission", the ACC and AHA probably in fact do intend to refer to outpatient encounters as well as inpatient ones.

Perhaps they should have stopped to define "admission", "visit", etc. first?  Well, if they had, it wouldn't have taken them long.  They could have adopted the definitions of these terms provided by the Ontology for General Medical Science, available here.

So we see that persons who are experts in a particular field of study, although necessary to the process of defining the terms they use, are not sufficient.

They should have consulted a good ontologist!

Standardizing Ontology a Top Priority for NICHD

The Eunice Kennedy Shriver National Institute for Child Health and Human Development held a series of "vision workshops" to create its vision for the next decade.

In a slide deck here (warning: PDF), the number one recommendation under the heading of "Transdisciplinary Science" is ...Standardize ontology, nomenclature, and data standards across disciplines.

The Director of NICHD, Dr. Alan Guttmacher, reitierated that recommendation today in a presentation to the Clinical and Translational Science Award Consortium Child Health Oversight Committee (CC-CHOC). The CC-CHOC meeting had a theme of "Quality Data from Pediatric Trials".

Speakers throughout the day emphasized the need for development and use of standards for collection of data in pediatric research.

Workshop and Tutorial Summary

I attended the Representing Adverse Events Workshop, and presented as part of the Improving Structured EHR Data Tutorial.

The Adverse Events Workshop confronted the issue of what exactly an adverse event is. A diversity of viewpoints frustrated attempts to reach consensus. The need for an adverse events ontology was called into question by some participants (and one non-participant with whom I discussed the issue afterwards). The concern is that various symptoms and diseases will be duplicated. For example, "fever" vs. "fever adverse event" and "rash" vs. "rash adverse event".

The outcome was an agreement to review use cases and to create a Google Group where participants can post links to their ontologies with requests for feedback. One participant took the action item of reviewing various artifacts (ontology and otherwise) that give definitions of 'adverse event' to look for commonalities.

The Improving Structured EHR Data Tutorial was given by Werner Ceusters and myself. As one of the instructors, I do not want to overstate the success of the tutorial, but there were excellent discussions about Referent Tracking and the advantages it has for making EHR data unambiguous. At least one participant approached me to express her appreciation for the potential of Referent Tracking, and her willingness to collaborate on projects in the future.

The hard work of improving our ability to leverage information to accelerate improvements in the quality and efficiency of biomedical research and patient care continues...

On the eve of ICBO 2011

I am here in Buffalo on the eve of the International Conference on Biomedical Ontology 2011. The conference begins with workshops and tutorials on Tuesday and Wednesday, followed by the 3-day conference program on Thursday, Friday, and Saturday.

The conference web site is here.

I will be attending the Representing Adverse Events Workshop tomorrow, then participating in a more meaningful way in the Improving Structured EHR Data Tutorial on Wednesday morning.

Over the next few days, it is my intent to post summaries of the best and worst of this year's ICBO, as I did at the last ICBO in 2009.

ISO TR 12310 Draft Standard - Problems and Issues

A draft of the ISO TR 12310 standard came across our desks in the last day or two, and we were surprised at the substantial problems that exist in its fundamental components. The title of the standard is "Health informatics — Principles and guidelines for the measurement of conformance in the implementation of terminological systems".

First, it defines a concept as "A concept is a single mental representation of some real or abstract thing."

This definition is problematic, because never in the history of terminological systems has anyone elucidated a theory of a basic unit of mental representation. What is a single mental representation? Or a unit of thought (as concept has also been defined)?

If I'm thinking of a car, is that a unit of thought? What if I'm thinking about a particular red, C-class Mercedes Benz with a particular Vehicle Identification Number? Or what if I'm just thinking about red, C-class Mercedes owned by stock brokers and that need new brakes and on which a tree fell during a Spring storm, in general? How many unique mental representations are there? One, two, three? How do we count unique ideas and single mental representations?

This question is not trivial, as the answer determines what concept representations we are permitted to have. Historically, the answer has been that anything anyone wants to put in the system is allowable. For a real-world example, consider the following "concept" from SNOMED CT: Family history of myocardial infarct in first degree female relative less than 65 years of age (situation).

The draft standard then goes on to say: "Concepts are should be unique within a code system." The grammatical problem aside, the standard is now engaged in use-mention confusion. For your mental representations are not part of any code system! The standard is using the word concept for both representations of concepts and concepts (themselves mental representations).

The draft standard then goes on to say: "A concept representation is a mechanism by which the system can express a concept." So it should be concept representations that are unique within a code system, not concepts.

But then the standard says: "Most code systems support multiple representations for each concept, sometimes even multiple representations of a given type."

How do code systems support interoperability, then?

Next, we are told that a "Concept id" is "A concept representation that is unique within the code system and that is used internally by the code system when referencing concepts."

Does that mean we cannot use concept ids outside the code system, either because it's impossible or disallowed? Surely some software applications are using SNOMED CT concept ids external to SNOMED CT itself? Are they out of conformance?

And that covers just pages 1 and 2.

NSF Funds Phenotype Ontology Coordination Network

The National Science Foundation recently awarded just over $500,000 to the University of South Dakota (USD) for the creation of the Phenotype Ontology Research Coordination Network. According to the grant proposal, the PI at USD is Dr. Paula Mabee and coPIs include Dr. Andrew Dean (North Carolina State University), Dr. Eva Huala (Stanford University), Suzanna Lewis (Lawrence Berkley National Laboratory), and Dr. Anne Maglia (Missouri University of Science and Technology).

The goal of the grant is to ...establish a network of scientists who are developing phenotype ontologies and to use this network to enable and enhance the research of all of those involved.

The specific aims are:

1. Develop anatomical reference ontologies for three key taxonomic clades;
2. Align and synchronize anatomical ontologies using homology and various types of similarity relations;
3. Define, test and document anatomy ontology development best practices and standards;
4. Reach out to ancillary phenotype groups to share with them common concepts and practices;
5. Educate the community about the methods for developing ontologies and their importance and utility in research.

The principals of the grant are soliciting participation in the network. To join, you can sign up here.

The Statistical Core Vocabulary: Can Definitions Really be this Bad?

I just learned of the Statistical Core Vocabulary (abbreviated as SCOVO), the latest version of which one may find here. This vocabulary defines itself as: A vocabulary for representing statistical data on the Web.

It contains 3 classes and 5 properties. For a small artifact that aims to capture the core entities in a domain, one ought to expect that careful attention would be paid to the definitions of the classes and properties, on the part of its creator(s).

Clearly, that is not the case here.

The SCOVO definition of Dataset is: a statistical dataset. This is equivalent to defining subatomic particle as: a positively-charged subatomic particle.

Similarly, the SCOVO definition of Item is: a statistical data item. This definition is even worse, being the equivalent of defining book as: a red hardcover book.

Finally, the SCOVO definition of its third and final class, Dimension, is: a dimension of a statistical data item. This definition is the equivalent of defining length as the length of a rectangular wooden table.

The situation for the properties in SCOVO is no better. Two of the properties even have the same name as two of the classes: dimension and dataset. The dataset relation is intended to link an item to a dataset. The dataset relation is thus the equivalent of having a relation liver to link the liver itself to the organism of which it is part.

By contrast, the Information Artifact Ontology (IAO) has taken pains to define the terms data item and dataset well. Definitions of these terms, and other terms used in their definitions, follow:

Data item: an information content entity that is intended to be a truthful statement about something (modulo, e.g., measurement precision or other systematic errors) and is constructed/acquired by a method which reliably tends to produce (approximately) truthful statements.

Information content entity: an entity that is generically dependent on some artifact and stands in relation of aboutness to some entity.

Note that for the definition of generic dependence, the IAO relies on a formal, upper ontology called Basic Formal Ontology.

Data set: a data item that is an aggregate of other data items of the same type that have something in common. Averages and distributions can be determined for data sets.

In the IAO, data items stand in the "is about" relation to some entity. Thus, data are linked to the world that they are about.

SCOVO does have one thing going for it: it serves as an exemplar of the worst possible way to define terms.