 Next speaker is Jeff Greta, who probably doesn't need an introduction to most people in this crowd. Jeff has been associated with various aspects of neuroinformatics for many years, starting with the FMRI data center at Dartmouth, which was a long while back, but more recently with the Neuroscience Information Framework and Intellex and Cycrunch, and he's going to talk a little bit about Intellex. Yeah, so thank you, and what I'm going to be doing today is giving a very brief overview of Intellex. So this is a new infrastructure component that we've been developing, and I'll try to tie into some of the fair components that Michelle actually talked about, because I think that's important. For those of you that actually want to see it in operation, come by tomorrow night. That's part of the demo session, poster session number 76. And so one of the things Michelle mentioned was standards, so there are a lot of standards out there. BioPortal, which is one of the places infrastructure sites for hosting ontologies, great site. You go there, you look and there are 722 ontologies. So you put in your term and you try to figure out, okay, which of these 20 ontologies do I choose? And so it gets a little unwieldy in terms of standards, in terms of vocabularies. So one of the things that we've done as part of the NIF project was within the neuroscience community, was try to bring together the various sort of domain ontologies that meet certain criteria. So using uberon for brain regions, cell ontology, gene ontology, working with gene ontology to incorporate some of the terms that we've added. So we've tried to sort of compose an ontology from these community ontologies, again, not doing the I have 14 standards, let me build a 15th, but try to take what's out there and bring it together. And so this is an activity that started early on in NIF and continues to this day. But the issue was how do we get actual users to get into these terms and vocabularies? The ontology is not something they could use. And so early on in NIF, we developed Neuralex, Stephen Larson and Mary Ann Martone sort of launched that a while ago, but it had some constraints. It was built on semantic media wiki. And to be able to create some of the tooling around it that we wanted to, it wasn't quite amenable to that. And so what we've done is we've basically ported Neuralex to this new platform that we're calling Interlex. And basically provides this community lexicon, so in a sense sort of the F, the findable for Michelle, because one of the issues as well is how do people find, you know, there is bioportal. But then there are also terms. So working with Reapernim in terms of some medical imaging, you know, the full ontology of DICOM labels, you know, is not really out there, right? So, you know, within the group, Carl Helmer and others, you know, developed sort of that vocabulary. That's something that we included in Interlex, right? So we're trying to bring together, you know, a neuroscience focused site where these terminologies and ontologies can be brought together, always of course linking out to the original terminologies and vocabularies. So it is an infrastructure component that is merged with a couple other utilities. So I mentioned the NIF ontology, you know, that actually feeds in Interlex and actually Interlex also feeds back into the ontology. We've also ported the NIF ontology completely into GitHub, so if anyone wants to make any comments or suggestions, you know, you can actually, you know, use your standard Git mechanisms to provide feedback and provide content. And we've also, a number of years ago, you know, developed with Chris Mungle and others, SciGraph, which is a Neo4j graph database service on top of ontologies. And so that's something else that we use in our infrastructure. So Interlex provides a terminology dashboard, you know, you can go in, you know, it provides a sort of user-friendly view of some of the components of the ontology. So if you want to see, you know, what children are available for Neuron, you know, then you can go into, you know, Prokinje cell, you know, it provides all the standard information in terms of synonyms and all that, as well as, you know, providing on the right there some other identifiers, so other identifiers from other ontologies and vocabularies from where this was found. In terms of working with the community, though, one of the important aspects is that, you know, when we're working, you know, in, you know, very scientific disciplines, we come up, you know, with terms that don't exist in ontologies and that's always been a very thorny issue as to how researchers can actually create a term, link it into other terminologies that exist as what FAIR, you know, sort of requires us to do, right, try to make it interoperable with what's already out there, you know, and a lot of times working with ontologies, you know, it's a months-long process, right, but if you want to sort of quickly get started and make a contribution, make a suggestion, what we do is provide an easy way for people to contribute terms into this coordinated set of vocabularies so that it is something that can be used by the researchers in the tooling that they're developing, you know, to test it out rather than, you know, these longer wait periods. And we also try to promote reuse, so whenever someone is trying to do something and add something, we actually always check to make sure, you know, are you trying to, you know, add something that's already been there. It does support, you know, an easy addition of relationships to other terms so you can go in, you can actually create, you know, relationships between terms that you enter or terms that already exist and you can also put in novel relationships. So again, we're trying to figure out a way to help crowdsource sort of the neuroscience-specific information around these terms. So for example, cell ontology has a lot of information about cells, but not a lot of information on how those cells, for example, relate to the anatomy when it comes to neuroscience or how they relate to neurotransmitters. And so this is the type of knowledge that we want to capture within the community. So this is the community knowledge on top of the base vocabularies and ontologies that sit there. And we do the same thing with annotations, right, so that if you have some informational artifact about one of these terms, you know, that can be added as well. So it's not just relationships between terms, but also these informational artifacts. Lastly, you know, one of the issues that we deal with in NIF is how do we know, you know, what someone says in a data source, you know, is related to a certain term. So, you know, what is BA3, right? Most people, if they're near imaging, would think, oh, Broadman Area 3, right? But what happens if it's B3? Is that vitamin B3, or is it, you know, Broadman Area 3? And so the other thing that we've been building into the system is the ability to crowdsource these term mappings so that people can submit information saying that this data source, you know, has this text in it, and that actually represents, you know, this term, you know, in the terminology vocabulary. And that's something that I think, you know, is important. It's something that's done many, many times, you know, by many different people. But, you know, trying to sort of aggregate this information, I think, across the community is important, you know, and to make that information publicly available. The last thing I'll talk about real briefly is that in addition to terms, a lot of what we're dealing with as well is common data elements, especially when we're trying to find data that exists. And common data elements are actually not so common. You know, you go to different repositories for common data elements, and there are hundreds of thousands of them, and sometimes they're duplicate, sometimes, you know, they represent the same thing. So, you know, I don't know, you know, how many different instruments there are to, you know, check handedness, but I know that they're more than 12, you know, but there are probably many more than that. And so one of the things that we have been working with, and this is, you know, specifically in the context of the ReapRNM project, is, you know, how do we actually federate those common data elements and using Interlex as a tool that sits behind that? So, for example, this is an example of something that, you know, is common in neuroimaging, you know, it's volumetric measurements of brain structures, and so there's actually, last time I was in Montreal, there was a meeting hosted by JB, and one of the things that came out of that was a metadata working group that's looking at how one can harmonize potential common data elements in the context, you know, of neuroimaging and do it in a practical way. And so one of the big issues is this discovery aspect around common data elements, and so one of the things that we're trying to figure out is how does one relate, for example, you know, there's an L caudate thick and there's a caudate left volume and there might be a CN underscore, you know, V somewhere else and some other instrument that all measure caudate thickness, and these are actual CDEs from the NIMH data archive from their data vocabulary, and so what we need to do to be able to discover this data is we need to actually have that mapping that says, okay, both, you know, the caudate, you know, L caudate thick and caudate left volume are both volume measurements of caudate, right, so we need to have that representation, and so what we've been doing is, you know, also incorporating these sort of federating elements within interlex, and we're also doing that so that they're within, that they have some context, right, so that we can use the rest of the ontology to actually do more refined searches so that, you know, when we do actually have this element that links together these individual CDEs around caudate volume, we know that that is actually a measurement about the caudate nucleus and it's linked into the terminology, right, and again, this gets to this whole concept, again, of interoperability and fair, you know, not building a silo but trying to incorporate these new elements into the context of what's already there so that you can actually do sort of semantically enhanced queries, so you could actually find then the volume of subcortical structure, you know, find me all data about volume of subcortical structures, right, and then you would be able to find all the common data elements that relate to that. And then in my last 20 seconds that I have, probably. Also, information on developer and tool support, so, you know, this is one thing that we're working on, so we do have APIs, you know, that interact with this. There's some new Python libraries that are coming out that we're interacting with Dave Kedron in context of re-pronim in terms of interacting with these terminologies programmatically as well. And so that's my brief overview. I guess I have time for maybe one question, and if not, come by the demo tomorrow. There's a question up there. Hi. I'm wondering about your, I guess I'll call it governance model for maintaining these ontologies and how you ensure that, and you're crowdsourcing fairly technical things, so how do you ensure that your crowdsourced definitions match to what the domain experts are saying when they use these terms in your ontologies? Yeah, so I mean, right now what we're, it's really in terms of the communities that are using the sort of the terms that are being generated. So it's, in a sense, an internal, community internal type vetting of those terms. You know, when we have gone and worked to bring sort of ontologies into other ontologies, so Mary Ann had worked, you know, a long time on subcellular anatomy descriptions that were taken in gene ontology, then there's sort of this additional step, you know, of, you know, doing, you know, more alignment and vetting, you know. But right now, the governance is really based, you know, on the, you know, the individual communities that are working to create sort of their term spaces. But, you know, Tom might have some other. I would just say that you want to take governance out of the picture entirely for these by giving communities the power to control their own vocabularies. That's a little bit. While they are actually also able to coordinate those definitions. So if I have a CDE, and in my domain, I talk about it in one way, but it's the same set of measurements that another domain talks about, that we shouldn't have to have a governance conversation about that. We should be able to look at how we actually talk about that and compare it directly between the communities. I think that's a good question we can post to the panel a little later. That's a long discussion.