so meets rnao karen eilbeck university of utah rnao consortium meeting may 28-29 2007

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

SO meets RNAO

Karen Eilbeck

University of Utah

RNAO Consortium Meeting

May 28-29 2007

• What SO is.

• How SO is used

• How SO is managed

• Where do SO and RNAO meet

• How SO and RNAO can work together

• If we have time - a demo of OBO-Edit

The Sequence Ontology describes the features of

biological sequence• Genome sequence

• Annotation of regions

• Coordinates

• Need to agree on meaning of terms. E.g. Does the CDS include the stop codon?

An annotation captures what we know about a gene

Ann

otat

ions

evid

ence

3 Alternate transcripts of Glut1 gene

5’ UTRStart codon

Coding exon

Transposon within intron

Structure of the ontology• SO is structured into a directed acyclic

graph. transcript exon

processed transcript

primary transcript

intron

clip

splice site

polyA site

protein coding primary transcript

nc primary transcript

mRNAncRNA CDS

UTR

PP

P

P

P

P

P

i

P

ii

ii

five_prime_UTR three_prime_UTR

i itRNA rRNA

ii

i

d

GFF3• SO is used to ‘type’ the features and

relationships.

Id type start end strand attributesctg123 . gene 1000 9000 . + . ID=gene00001;Name=EDEN ctg123 . TF_binding_site 1000 1012 . + . ID=tfbs00001;Parent=gene00001 ctg123 . mRNA 1050 9000 . + . ID=mRNA00001;Parent=gene00001;Name=EDEN.1 ctg123 . mRNA 1050 9000 . + . ID=mRNA00002;Parent=gene00001;Name=EDEN.2 ctg123 . mRNA 1300 9000 . + . ID=mRNA00003;Parent=gene00001;Name=EDEN.3 ctg123 . exon 1300 1500 . + . ID=exon00001;Parent=mRNA00003 ctg123 . exon 1050 1500 . + . ID=exon00002;Parent=mRNA00001,mRNA00002ctg123 . exon 3000 3902 . + . ID=exon00003;Parent=mRNA00001,mRNA00003 ctg123 . exon 5000 5500 . + . ID=exon00004;Parent=mRNA00001,mRNA00002,mRNA00003 ctg123 . exon 7000 9000 . + . ID=exon00005;Parent=mRNA00001,mRNA00002,mRNA00003

relationshipsterms

Why we made SO

• Standardize vocabulary used in genomics.

• Clarify the relationships between the terms.

• Make genomics data more computable by adding semantics to the sequence. Its not just about sequence similarity.

What is the scope of SO?

• Features that can be located on a sequence with coordinates. exon, promoter, binding_site

• Properties of these features:– Sequence attributes

• Maternally_imprinted

– Consequences of mutation• mutation_affecting_editing

– Chromosome variation• aneuploid

The SO community

• Model Organism DB– SGD– (MGI)– FlyBase– WormBase– DictyBase– Pombe

• GMOD• Comparative

genomics• MGED Ontology• NLP

Genome annotation unification

• The model organism databases use SO to type their features.

• The GFF3 file format for annotation, the Chado db schema and DAS2 annotation protocol rely on SO to type features.

Genomic analysis

• The Comparative Genomics Library written in Perl uses SO based annotations to perform complex analysis over multiple genomes.– Yandell M, Mungall CJ, Smith C, Prochnik S,

Kaminker J, Hartzell G, Lewis S, Rubin GM. 2006. Large-Scale Trends in the Evolution of Gene Structures within 11 Animal Genomes. PLoS Comput Biol. 2:e15

Genome data integration

• Multiple genomes are organized using SO:– Flymine, – Gramene, – the BRCs

NLP/text mining

• Recently SO have been used for some new projects - – Semantic enrichment by the Royal Society

of Chemistry.– Anaphora resolution by the NLIP group in

Cambridge.

How SO is managed

• SO uses CVS to manage and version the ontology.

• There is a mailing list for developers to get things off their chest.

• There is a tracker for term suggestions• There are workshops when we get a

critical mass for a given problem. We want to do more workshops.

• SO is expressed in OBO format.

Example of OBO format• http://www.geneontology.org/GO.format.obo-1_2.sht

ml

[Term]id: SO:0000587name: group_I_introndef: "Group I catalytic introns are large self-splicing ribozymes. They catalyse their own excision from mRNA, tRNA and rRNA precursors in a wide range of organisms. The core secondary structure consists of 9 paired regions (P1-P9). These fold to essentially two domains, the P4-P6 domain (formed from the stacking of P5, P4, P6 and P6a helices) and the P3-P9 domain (formed from the P8, P3, P7 and P9 helices). Group I catalytic introns often have long ORFs inserted in loop regions." [http://www.sanger.ac.uk/cgi-bin/Rfam/getacc?RF00028]subset: SOFAis_a: SO:0000188 ! intron

OBO and OWL

• http://purl.org/obo/owl/SO

• Mapping OBO and OWL http://www.bioontology.org/wiki/index.php/OboInOwl:Main_Page

Navigate SO using OBO-Edit

Structure of the ontology

Search the ontology

Details for selected term

All parents of the term

Annotating with SO and RNAO

The nanos translational control element represses translation in somatic cells by a Bearded box-like motif. ･ Duchow HK, Brechbiel JL, Chatterjee S, Gavis ER. Developmental Biology Volume 282, Issue 1, 1 June 2005, Pages 207-217

• AGAGGGCGAATCCAGCTCTGGAGCAGAGGCTCTGGCAGCTTTTGCAGCGTTTATATAACATGAAATATATATACGCATTCCGATCAAAGCTGGGTTAACCAGATAGATAGATAGTAACGTTTAAATAGCGCCTGGCGCGTTCGATTTTAAAGAGATTTAGAGCGTTATCCCGTGCCTATAGATCTTATAGTATAGACAACGAACGATCACTCAAATCCAAGTCAATAATTCAAGAATTTATGTCTGTTTCTGTGAAAGGGAAACTAATTTTGTTAAAGAAGACTTACAATATCGTAATACTTGTTCAATCGTCGTGGCCGATAGAAATATCTTACAATCCGAAAGTTGATGAATGGAATTGGTCTGCAACTGGTCGCCTTCATTTCGTAAAATGTTCGCTTGCGGCCGAAAAATTTCGATATATCTACAATTGATCTACAATCTTTACTAAATTTTGAAAAAGGAACACTTTGAATTTCGAACTGTCAATCGTATCATTAGAATTTAATCTAAATTTAAATCTTGCTAAAGGAAATAGCAAGGAACACTTTCGTCGTCGGCTACGCATTCATTGTAAAATTTTAAATTTTGACATTCCGCACTTTTTGATAGATAAGCGAAGAGTATTTTTATTACATGTATCGCAAGTATTCATTTCAACACACATATCTATATATATATATATATATATATATATATATATATATATATATGTTATATATTTATTCAATTTTGTTTACCATTGATCAATTTTTCACACATGAAACAACCGCCAGCATTATATAATTTTTTTATTTTTTTAAAAAATGTGTACACATATTCTGAAAATGAAAAATTCAATGGCTCGAGTGCCAAATAAAGAAATGGTTACAATTTAAGG

Translational control element

Overlap with RNAO

• SO provides regions of sequence - start and stop coordinates with regards to the whole sequence - i.e. assembly / chromosome– Transcripts and parts of transcripts– Some secondary structure – Some motifs– Results of algorithms such as blast

SO names features

Secondary structure

• This part of SO needs work.

• Any volunteers?

Divergent from RNAO

• Where do SO and RNAO differ dramatically?– Multiple sequence alignments. SO does

not provide a solution to this. It does however provide the terms to describe the results of sequence similarity searches.

– Numerical results. SO has not needed to use values so far.

RNAO working groups

• Motif identification/annotation

• RNA interaction

• Biochemical-structure mapping

• Multiple sequence alignment

• Backbone conformation

• Base stacking

Working together

• Remain 2 separate ontologies.

• Give SO annotators option of ‘importing’ RNAO terms using the OBO programs

• SO and RNAO work together to align key terms in their ontologies.

SO is still evolving

• RNAO could use the SO features to describe regions of sequence

• SO could reference RNAO for detailed annotation of structure and biochemical features.

Multiple ontologies in OBO

• 2 options.1. The ontologies reference each other:

• Will always need to load both ontologies

2. There is a mapping file that you can load to import external terms.• Maintain separate ontologies and keep

mapping up to date.

http://obofoundry.org/wiki/index.php/Mappings

Example: Importing terms from SCOR.

• 1. Made an OBO file from a subset of SCOR terms

• 2. Work out where there is overlap

• 3. Make OBO mapping file between the two ontologies

• 4. Load all 3 files at once.

format-version: 1.2date: 16:05:2007 15:26saved-by: kareneilbeckauto-generated-by: OBO-Edit 1.100

[Term]id: SC:0000000name: hairpin_loop

[Term]id: SC:0000001name: diloopis_a: SC:0000000 ! hairpin_loop

[Term]id: SC:0000002name: triloopis_a: SC:0000000 ! hairpin_loop

…

format-version: 1.2date: 24:05:2007 10:37saved-by: kareneilbeck

import: so-xp.oboimport: scor2.obo

id: SC:0000015 hairpin loopis_a: SO:0000715 is_a RNA motif

id: SC:0000016 internal loopis_a: SO:0000715 is_a RNA motif

id: SC:0000035 tertiary interactionis_a: SO:0000122 is_a RNA sequence

secondary structure

scor.obo mapping file

OBO-Edit DEMO

• Fingers crossed…

Possible action items

• A SO-RNAO mailing list for discussion of collaboration

• Phone/skype/webinars at intervals to keep track of progress.

• GFF3 http://www.sequenceontology.org/gff3.shtml

• Apollo http://www.fruitfly.org/annot/apollo/

• SO http://www.sequenceontology.org

• OBO-Edit http://sourceforge.net/projects/geneontology

• OBO foundry http://www.obofoundry.org

• GO-perl http://www.godatabase.org/dev/go-perl/doc/go-perl-doc.html

Resources

Acknowledgements• SO is funded as part fo the Gene

Ontology Consortium, via the NIH P41-HG002274

• People:– Suzi Lewis and Michael Ashburner - the vision– Chris Mungall - programming infrastructure– John Richter - made OBO-Edit

• keilbeck@genetics.utah.edu