Difference between revisions of "MGI"
|Line 159:||Line 159:|
Literature curation continues to be the major focus of our annotation efforts. We continue to explore natural language processing tools to aid in identifying papers
Literature curation continues to be the major focus of our annotation efforts. We continue to explore natural language processing tools to aid in identifying papers .
''Computational annotation strategies:''
''Computational annotation strategies:''
As always current strategies involve use of translation table to mine SwissProt
As always current strategies involve use of translation table to mine SwissProt InterPro domainsfor IEA annotation. These are performed automatically on a nightly basis and require little human intervention.
Revision as of 15:47, 14 September 2010
Mouse Genome Informatics Summary, 2010
lasted edited 9-14-10 by hjd
Harold J Drabkin*
- Funded entirely or partially by GO
We continue to put emphasis on those genes selected for the Reference Genome Project. Additional emphasis has been placed on certain genes associated with lung development.
|Annotation Type||09_Dec_10||09_Dec_09||Change||% Change|
|Total Genes annotated (with at least one GO term of any kind):||
|Total Manual Annotation|
|Number of Genes||
|SwissProt to GO||
|Interpro to GO||
|EC to GO||
|* 100% of current gene models|
The large changes in manual (non-IEA) curated genes are the result of the addition of 5 efforts.
- Annotating genes with no literature annotation to one or more of the three roots. Quality control reports now alert curators for any genes annotated to one or more of the three roots when new literature is added to MGI and the literature is indexed to these genes.
- We have added a pipeline to create annotations inferred by sequence orthology (ISO) to mouse genes that have rat orthologs where the rat orthologs have been annotated by experiment.
- We have added a pipeline to load GOA annotations to mouse genes.
- We have added a pipeline to add annotations based on intra-ontology relationships.
- We have added a pipeline to add annotations by on sequence similarity (ISS)based on Panther family membership via the Reference Genome projet.
Methods and strategies for annotation
Literature curation continues to be the major focus of our annotation efforts. We continue to explore natural language processing tools to aid in identifying the genes in papers selected for GO.
Computational annotation strategies:
As always current strategies involve use of translation table to mine SwissProt keywords, InterPro domains, and EC numbers for IEA annotation. These are performed automatically on a nightly basis and require little human intervention.
Priorities for annotation
- Genes assigned by Reference Genome Project (everyone)
- Isoform curation (Harold, Protein Ontology project); now co-ordinating with 1 by focusing on reference genes that have isoforms.
- Genes with no GO annotation but with literature (Li and Dmitry)
- Genes with only IEA annotation but with literature (Li)
- Genes identified as being important in lung development (Dmitry)
- Genes marked as having GO annotation completed, but now having new literature (Dmitry)
Presentations and Publications
A. Ontology Development Contributions:
1. David Hill has worked on a team with Tanya Berardini, Chris Mungall, Midori Harris and Jane Lomax to develop cross-products within and among the three GO namespaces.
2. David Hill continues to work with Chris Mungall and Tanya Berardini to inter-ontology links in GO.
3. David Hill has worked with Yasmin Alam-Faruque, Doug Howe, Midori Harris, Susan Tweedie, Becky Foulger and community experts to expand the kidney development portion of GO.
4. Alexander Diehl is active in the Signaling GO content development group.
5. Alexander Diehl is active in the Virus Term GO content development group.
6. Alexander Diehl continues to act as the GOC liaison to the Infectious Disease Ontology and Vaccine Ontology groups and to act on term requests for the GO from those groups.
7. Alexander Diehl and Chris Mungall are leading a project to revise the Cell Ontology and increase its utility as a source for cell type-specific GO terms and in co-annotation with GO terms. This project is funded by an ARRA Competitive Revision to the main GO Consortium grant; funding began on September 30, 2009. Terrence Meehan at MGI was recruited to work as a full time curator on the Cell Ontology under the direction of Alexander Diehl. Our first large project on the CL is developing the hematopoietic cell type terms into a full logical-definition/cross product format.
8. Alexander Diehl and Judith Blake recently co-mentored Morgan V. Hightshoe, a member of the 2009 Jackson Laboratory Summer Studen.t Program, in a project to revise the representation of nervous system cell types in the Cells Ontology. This work is being continued by a high school intern, Wade Valleau, as part of our larger work on the cell ontology.
9. David Hill and Jane Lomax oversee the biological content development of GO. In particular, all new developmental biology-related terms submitted to SourceForge are handled by David Hill and all new 'regulation' terms are handled by David Hill and Tanya Berardini.
Annotation Outreach and User Advocacy Efforts:
- The Protein Ontology project continues to provide a web interface (http/pir.georgetown.edu/cgi-bin/pro/race_pro) whereby functional annotation using the GO can be applied to PRO submissions. These are reviewed by Cecilia Arighi of Georgetown. At present, only the PRO curators (Georgetown and MGI) are using the tool, but it is available to anyone.
- Harold is mentoring to curators in Donna Slonim's group at Tufts, Heather Wick and Craig Fournier. They are focusing on genes involved in human fetal development.
- David Hill continues to serve on the GO-help rota.
We are now suppling a GAF 2.0 format file to GOC with column 17 isoform data filled in. This file is also available directly from our own FTP site. Column 16 fill in, starting with specifying cell type will be implemented shortly.
Alexander Diehl co-wrote an ARRA Competitive Revision extension to the Gene Ontology Consortium grant HG000273, along with Chris Mungall and Judith Blake for the purpose of revising the Cell Ontology, and using the CL in conjunction with the GO in annotation and in cross-product term formation. This grant was funded on September 30, 2009.
Dmitry Sitnikov continues collaboration with Larry Hunter's group (Dr. Mike Bada) to establish a large, high-quality, corpora of full-text publications, expertly annotated with expressive knowledge representations, to improve the performance of a wide variety of biochemical text mining systems and to create new approaches to text mining. This involves systematically training and evaluating a broad sample of information extraction methods for key tasks, including concept and relationship identification. This effort can also be useful in synonym improvements to the GO between GO terms and certain biological concepts and terminology used in the literature. So far, more then 80 artictles of the chosen 98 have been annotated for molecular function and process. Additionally, Dmitry is working to improve the usefulless of MGI internal GO QC reports.
As the designated coordinator of the MGI/GO project with the GO Reference Genome project, Li Ni participates in annotations of genes assigned by the Reference Genome Project, maintain the mouse Reference Genome list on MGI GO wiki and Google spreadsheet, maintain the Reference Genome status table on GO wiki, oversees the curation of Reference Genome Genes for the mouse group.
Mary Dolan has been involved in a collaboration with Carol Bult at MGI on aligning gene ontology annotations for mouse genes assigned to MouseCyc pathways. See