Search

Single-cell RNA-Sequencing (scRNA-Seq) has undergone major technological advances in recent years, enabling the conception of various organism-level cell atlassing projects. With increasing numbers of datasets being deposited in public archives, there is a need to address the challenges of enabling the reproducibility of such data sets. Here, we describe guidelines for a minimum set of metadata to sufficiently describe scRNA-Seq experiments, ensuring reproducibility of data analyses.

A major goal of proteomics is the complete description of the protein interaction network underlying cell physiology. A large number of small scale and, more recently, large-scale experiments have contributed to expanding our understanding of the nature of the interaction network. However, the necessary data integration across experiments is currently hampered by the fragmentation of publicly available protein interaction data, which exists in different formats in databases, on authors' websites or sometimes only in print publications. Here, we propose a community standard data model for the representation and exchange of protein interaction data. This data model has been jointly developed by members of the Proteomics Standards Initiative (PSI), a work group of the Human Proteome Organization (HUPO), and is supported by major protein interaction data providers, in particular the Biomolecular Interaction Network Database (BIND), Cellzome (Heidelberg, Germany), the Database of Interacting Proteins (DIP), Dana Farber Cancer Institute (Boston, MA, USA), the Human Protein Reference Database (HPRD), Hybrigenics (Paris, France), the European Bioinformatics Institute's (EMBL-EBI, Hinxton, UK) IntAct, the Molecular Interactions (MINT, Rome, Italy) database, the Protein-Protein Interaction Database (PPID, Edinburgh, UK) and the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING, EMBL, Heidelberg, Germany).

To the Editor Although there is only one human genome sequence, different genes are expressed in many different cell types and tissues, as well as in different developmental stages or diseases. The structure of this ‘expression space’ is still largely unknown, as most transcriptomics experiments focus on sampling small regions. We have constructed a global gene expression map by integrating microarray data from 5,372 human samples representing 369 different cell and tissue types, disease states and cell lines. These have been compiled in an online resource (http://www.ebi.ac.uk/gxa/array/U133A) that allows the user to search for a gene of interest and find the conditions in which it is over- or underexpressed, or, conversely, to find which genes are over- or underexpressed in a particular condition. An analysis of the structure of the expression space reveals that it can be described by a small number of distinct expression profile classes and that the first three principal components of this space have biological interpretations. The hematopoietic system, solid tissues and incompletely differentiated cell types are arranged on the first principal axis; cell lines, neoplastic samples and nonneoplastic primary tissue–derived samples are on the second principal axis; and nervous system is separated from the rest of the samples on the third axis. We also show below that most cell lines cluster together rather than with their tissues of origin.

The Functional Genomics Experiment data model (FuGE) has been developed to facilitate convergence of data standards for high-throughput, comprehensive analyses in biology. FuGE models the components of an experimental activity that are common across different technologies, including protocols, samples and data. FuGE provides a foundation for describing entire laboratory workflows and for the development of new data formats. The Microarray Gene Expression Data society and the Proteomics Standards Initiative have committed to using FuGE as the basis for defining their respective standards, and other standards groups, including the Metabolomics Standards Initiative, are evaluating FuGE in their development efforts. Adoption of FuGE by multiple standards bodies will enable uniform reporting of common parts of functional genomics workflows, simplify data-integration efforts and ease the burden on researchers seeking to fulfill multiple minimum reporting requirements. Such advances are important for transparent data management and mining in functional genomics and systems biology.

To the editor: As researchers in bioinformatics turn increasingly to ontological tools and resources to help them to solve problems of information integration, data annotation, natural language processing and automated reasoning, the need for proven, useful ontologies in biomedical research will continue to grow exponentially. We are delighted to see Nature Biotechnology taking up this important issue, with the publication of a commentary on the MGED (Microarray Gene Expression Data Society) ontology in the September issue by Soldatova and King (Nat. Biotechnol. 23, 1095–1098, 2005). Soldatova and King make the important point that ontology builders need to be aware of emerging standards and best practices. When they trace the origins of modern work on ontology to Aristotle, however, it is important to remember that philosophers have been debating Aristotle’s categories continuously for more than two millennia. The number of ways in which we can slice up the world into categories is in practice unlimited, and it is not always obvious which distinctions one needs to make, for the purposes of the MGED ontology, or for any other purpose. Moreover, finding problems in virtually any extant ontology is a trivial exercise. Soldatova and King recommend the use of the Suggested Upper Merged Ontology (SUMO) as an emerging standard with the capacity to bring coherence to ontologies in the biomedical domain. Unfortunately, SUMO in its current form embodies no well-defined criteria to determine which classes should be properly included within its scope, with unfortunate consequences for its overall integrity and usability for purposes of ontology integration in the life sciences. Should an upper ontology designed for purposes of robust integration of biomedical ontologies include classes such as Monkey or BodyCovering? Or odd disjunctive classes, such as FruitOrVegetable? It is self-evident that methods to assist the empirical evaluation of both ontology content and structure are urgently required. As Soldatova and King themselves acknowledge, “the engineering of ontologies is still a relatively new research field.” Much of the most influential ontology work in biomedicine has been stimulated by the pressing needs of bench biologists themselves in managing burgeoning quantities of data. As a consequence, many of the ontologies developed thus far are somewhat unprincipled in comparison to what we now know can be achieved. Today, however, we have reached the point where an increasing number of biomedical scientists are recognizing the importance of learning about standards of good practice in ontology development and of adhering to those standards whenever possible.1 The newly created National Center for Biomedical Ontology2, formed under the US National Institutes of Health (NIH) roadmap, is a direct acknowledgment of this need. We are principal participants of this center and have a particular interest in improving the quality of all ontologies developed for use in biomedicine. The center will be attempting, through systematic outreach activity and through testing and dissemination of good ontology practices, to aid biomedical investigators in the construction of ontologies that adhere to proven conventions and knowledge-representation formalisms3. We will conduct workshops designed to promote collaboration among different groups of ontology developers and to assist biomedical researchers in developing and applying ontologies precisely tailored to their needs. We believe that the establishment of the center will offer an opportunity to enhance consistency and clarity in biomedical ontologies and to increase the prospects for their interoperability, for example, through the use of common, well-defined relationships along the lines applied to all new entries in the Open Biomedical Ontologies library4,5. The central role of good ontologies in biomedical informatics is unquestioned. What we need now is research to establish how best to achieve our broader goals through the formalization and integration of biomedical knowledge.