Neuron Phenotype Ontology – classifying the brain's building blocks

© 2022 EPFL

© 2022 EPFL

As the study of neuroscience accelerates ever faster and our understanding of the intricate workings of the brain deepens, the need to define a standardized approach to naming neuron types has become imperative. As the building blocks of the brain, neurons are studied intensely with vast amounts of data being generated and shared. With this comes several challenges as there is more and more data to describe different cell types. Furthermore, the literature contains many neuron names that are commonly used and accepted, even when it is unclear how such common usage types relate to the many proposed evidence-based types that are based on the results of new techniques. In addition, there is the significant question of comparing different data sets across labs. This all points to an urgent need for a standardized approach to naming neurons and for the organization of knowledge about their properties.

At the turn of the 19th century, the Spanish physician Santiago Ramón y Cajal mapped every detail of the forest of neurons that make up the brain using pen and paper, calling them the “butterflies of the brain”. At the time, he benefited from a newly discovered technique - the Golgi stain, to reveal neurons as individual entities of remarkably different shapes. Today, our knowledge of neuron types (as with cell types) has continued to evolve as new experimental techniques emerge and new technologies are enabling further characterization of additional dimensions, including some that may be foundational, so our concept of cell types is likely to evolve even further.

In parallel, large projects such as the EPFL Blue Brain Project, US Brain Initiative Cell Census Network, Human Cell Atlas and others, are generating vast amounts of data and characterizing large numbers of neurons throughout the nervous system. These projects have as one of their common end goals, the integration and analysis of large quantities of cellular data to derive new taxonomic classification of neurons across neural structures. The aim is to arrive at a new understanding of what constitutes a cell type in the nervous system. However, to manage this process, it is clear that a consistent naming scheme for neurons is needed, so that as new types are discovered, their findings can be reported and compared in an organized and useful way.

An open source standardized and automatable approach for naming cell types

In a paper recently published in Neuroinformatics, a collaboration of Scientists from the Swiss EPFL Blue Brain Project, University of California, San Diego, University of Toronto, Canada and the Krembil Centre for Neuroinformatics, Canada, propose an interoperable knowledge representation - Neuron Phenotype Ontology (NPO). The NPO is an open source standardized and automatable approach for naming cell types and normalizing their constituent phenotypes using identifiers from community ontologies as a common language. It is designed to enable scientists to discover which cell types (or potential cell types) share similar properties and to help scientists understand when the cell types they observe are the same or similar to other cell types described in the literature or from other laboratories. Moreover, it provides a framework for systematically organizing knowledge about cellular properties and enables interoperability with existing neuron naming schemes.

Mohameth François Sy, Data and Knowledge Engineering Manager at the Blue Brain, explains, “What is unique about the NPO, is that it allows us to communicate about and compare measured neuronal phenotypes in a way that reflects our understanding as humans but can also be fully managed using modern computational methods. The NPO and the associated knowledge environment provide a bridge between classifications generated using high throughput and integrative techniques and our accumulated knowledge over the past 100 years on cell types in the nervous system”.

To evaluate the NPO, a knowledge base was populated with three independent cortical neuron classifications derived from published data sets that describe neurons according to molecular, morphological, electrophysiological, and synaptic properties. Competency queries to this knowledge base demonstrate that the NPO knowledge model enables interoperability between the three test cases and neuron names commonly used in the literature.

Applying the FAIR data principles

“Traditionally, neuron type names and definitions evolved through discussions in the literature. Today, large-scale data-driven studies are identifying new ways of defining neuron types all the time,” discloses Prof. Sean Hill, Professor at the Blue Brain Project and the Scientific Director of the Krembil Centre for Neuroinformatics. “We need new ways of defining neuron types that can span the classical names from the literature and the new data-driven neuron types. The NPO provides a semantic data model to do just that, and implements the FAIR principles making these neuron types findable, accessible, interoperable and reusable. Thus, the NPO provides the foundation for large interoperable databases of neuron data and literature worldwide.”


For more information, please contact – Blue Brain Communications Manager – [email protected]

Citation - Gillespie, T. H., Tripathy, S. J., Sy, M. F., Martone, M. E., & Hill, S. L. (2022). The Neuron Phenotype Ontology: A FAIR approach to proposing and classifying neuronal types. Neuroinformatics.

Notes for Editors and Scientists

FAIR Properties of the NPO
The NPO was designed to be consistent with the FAIR principles. In table five of the paper, the authors show how the NPO achieves FAIR using the rubric in Hodson et al. (2018). The key features are machine readability, the use of identifiers (FAIR vocabularies), common knowledge representation languages and community standards.

Open Source – Open Science
The NPO is composed of two parts. A set of core ontology files that define a data model for neuron types, and the NPOKB, the collection of neuron types defined using the NPO core ontology data model. The NPO as well as all data and code referenced are available for reuse under open licenses - go to the Data and Code availability information in the paper.

Authors and Affiliations

Department of Neuroscience, University of California, San Diego - Thomas H. Gillespie & Maryann E. Martone
Department of Psychiatry, University of Toronto, Canada - Shreejoy J. Tripathy & Sean L. Hill
Department of Physiology, University of Toronto, Canada - Shreejoy J. Tripathy & Sean L. Hill
Krembil Centre for Neuroinformatics, Centre for Addiction and Mental Health, Toronto, Canada - Shreejoy J. Tripathy & Sean L. Hill
Blue Brain Project, École Polytechnique Fédérale de Lausanne (EPFL), Campus Biotech, 1202, Geneva, Switzerland - Mohameth François Sy & Sean L. Hill


This work was supported by NIH Brain Initiative award 5U24MH114827-04, a Canadian Institute for Health
Research post-doctoral fellowship and National Institutes of Health grant MH111099, the Krembil Foundation, and by funding to the Blue Brain Project, a research center of the École Polytechnique Fédérale de Lausanne (EPFL), from the Swiss government’s ETH Board of the Swiss Federal Institutes of Technology. Open access funding provided by EPFL Lausanne.