How Can Rdf (Resource Description Framework) "Semantic Web" Improve An Integrated Genome Viewer?

Question

[GreenDiamond]

A group recently decoded a coral Acropora digitifera. This species one of the most sensitive to rises in seawater temperature, and it is hoped that the genomic information can help scientists better understand coral death mechanisms from bleaching attributed to global warming. (Article in Nature: "Using the Acropora digitifera genome to understand coral responses to environmental change")

In their new article ("MarinegenomicsDB: An Integrated Genome Viewer for Community-Based Annotation of Genomes"), they introduce a database that allows for community-based annotation of genomes. While I could follow most of the paper, I would very grateful to better understand the final paragraph regarding their future directions:

"An attempt to integrate data across the entire web has been recommended by the World Wide Web Consortium (W3C; http://www.w3.org/2004/01/sws-pressrelease.html.en). Simply by following the resource description framework (RDF), which is a more suitable format for processing by machines than the current standard of HTML, the data stored at multiple websites can be automatically organized without human interpretation. To take advantage of this new generation of web technology, the “semantic web,” we will publish our data in RDF in the near future."

1) If you are interested and read this short paper (3 pages), what do you think this last paragraph means?

2) What steps would need to be done to "publish [their] data in RDF"?

3) How exactly would RDF change the functionality and efficiency of their database?

I have very little understanding of these software, so you need not worry about sounding patronizing to me; I just want a basic understanding of what that final paragraph means, and if you can answer any of my questions above for me, I would appreciate it!!

Thank you...

[krisjvr]

Isn't OWL based on RDF? It is just the next layer.

[DG]

RDF is a fairly broad description/standard. You're right in that OWL is a specific implementation of broader RDF standards.

[DG]

BTW you should really add these sorts of things as comments and not new answers to your question

[krisjvr]

@Dan .. "BTW you should really add these sorts of things as comments and not new answers to your question"

FYI

The "ADD COMMENT" in gray box below questions and the "ADD REPLY" links below responses are not active.

The "Add Answer" is the only text entry box I see. Let me know if I'm missing something. This is the only place I can enter text.

[DG]

That's very odd. It may be a combination of browser and browser add-ons (flashblock, noscript, adblock, etc) which sometimes deactivate some types of elements on web pages. I forget if there is a bug submission system for the webpage

Answer 1

[Pierre Lindenbaum]

"How can RDF (resource description framework) "semantic web" improve an integrated genome viewer?"

if your genome browser displays a raster image (PNG, JPG) one could imagine to insert RDF meta-data in the image to describe it ( e.g;

uri:myview my:has_build ucsc:hg19 .
uri:myview my:chrom ucsc:hg19:chr1 .
hg19:chr1 my:length "250000000" .
uri:myview my:start "12345" .
uri:myview my:end "5678" .
uri:myview my:track snp
uri:myview my:track gene

if your genome browser displays a vectorial image (SVG) (and it would be already a big file) one could imagine to insert RDF meta-data in the image to describe each item and any RDF engine could use that information. Say you're displaying a rectangle for a gene:

<svg:rect x="1" y="1" with="123" height="100" title="erotonin receptor 1B " rdf:about="&lt;a href=" http:="" bio2rdf.org="" rdfxml="" geneid:37191"="" rel="nofollow">http://bio2rdf.org/rdfxml/geneid:37191" />

from this simple uri, one can retrieve all the linked data to that gene:

http://bio2rdf.org/rdfxml/geneid:37191

<rdf:RDF xmlns:rdfs="&lt;a href=" http:="" www.w3.org="" 2000="" 01="" rdf-schema#"="" rel="nofollow">http://www.w3.org/2000/01/rdf-schema#"
    xmlns:ns0pred="http://purl.org/dc/terms/"
    xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">

<rdf:Description rdf:about="&lt;a href=" http:="" bio2rdf.org="" geneid:37191"="" rel="nofollow">http://bio2rdf.org/geneid:37191">
    <rdf:type rdf:resource="&lt;a href=" http:="" bio2rdf.org="" geneid:vocabulary:protein-coding-gene"="" rel="nofollow">http://bio2rdf.org/geneid:vocabulary:protein-coding-gene"/>
    <rdf:type rdf:resource="&lt;a href=" http:="" bio2rdf.org="" geneid:vocabulary:Gene"="" rel="nofollow">http://bio2rdf.org/geneid:vocabulary:Gene"/>
    <rdfs:label>Serotonin receptor 1B (symbol:5-HT1B,taxon:7227) [geneid:37191]</rdfs:label>
    <has_taxid xmlns="&lt;a href=" http:="" bio2rdf.org="" geneid_vocabulary:"="" rel="nofollow">http://bio2rdf.org/geneid_vocabulary:" rdf:resource="http://bio2rdf.org/taxon:7227"/>
    <has_ensembl_gene_identifier xmlns="&lt;a href=" http:="" bio2rdf.org="" geneid_vocabulary:"="" rel="nofollow">http://bio2rdf.org/geneid_vocabulary:" rdf:resource="http://bio2rdf.org/ensembl:FBgn0263116"/>
    <has_ensembl_protein_identifier xmlns="&lt;a href=" http:="" bio2rdf.org="" geneid_vocabulary:"="" rel="nofollow">http://bio2rdf.org/geneid_vocabulary:" rdf:resource="http://bio2rdf.org/ensembl:FBpp0085783"/>
    <has_ensembl_protein_identifier xmlns="&lt;a href=" http:="" bio2rdf.org="" geneid_vocabulary:"="" rel="nofollow">http://bio2rdf.org/geneid_vocabulary:" rdf:resource="http://bio2rdf.org/ensembl:FBpp0289709"/>
    <has_rna_ensemble_identifier xmlns="&lt;a href=" http:="" bio2rdf.org="" geneid_vocabulary:"="" rel="nofollow">http://bio2rdf.org/geneid_vocabulary:" rdf:resource="http://bio2rdf.org/ensembl:FBtr0300482"/>
    <has_rna_ensemble_identifier xmlns="&lt;a href=" http:="" bio2rdf.org="" geneid_vocabulary:"="" rel="nofollow">http://bio2rdf.org/geneid_vocabulary:" rdf:resource="http://bio2rdf.org/ensembl:FBtr0086599"/>
    <has_protein_accession xmlns="&lt;a href=" http:="" bio2rdf.org="" geneid_vocabulary:"="" rel="nofollow">http://bio2rdf.org/geneid_vocabulary:" rdf:resource="http://bio2rdf.org/refseq:NP_001163201.1"/>
    <has_protein_accession xmlns="&lt;a href=" http:="" bio2rdf.org="" geneid_vocabulary:"="" rel="nofollow">http://bio2rdf.org/geneid_vocabulary:" rdf:resource="http://bio2rdf.org/refseq:78707343"/>
    <has_protein_accession xmlns="&lt;a href=" http:="" bio2rdf.org="" geneid_vocabulary:"="" rel="nofollow">http://bio2rdf.org/geneid_vocabulary:" rdf:resource="http://bio2rdf.org/genbank:AAF57610.4"/>
    <has_protein_accession xmlns="&lt;a href=" http:="" bio2rdf.org="" geneid_vocabulary:"="" rel="nofollow">http://bio2rdf.org/geneid_vocabulary:" rdf:resource="http://bio2rdf.org/genbank:ACZ94473.1"/>
    <has_protein_accession xmlns="&lt;a href=" http:="" bio2rdf.org="" geneid_vocabulary:"="" rel="nofollow">http://bio2rdf.org/geneid_vocabulary:" rdf:resource="http://bio2rdf.org/genbank:ACN86076.1"/>
    <has_protein_accession xmlns="&lt;a href=" http:="" bio2rdf.org="" geneid_vocabulary:"="" rel="nofollow">http://bio2rdf.org/geneid_vocabulary:" rdf:resource="http://bio2rdf.org/genbank:P28286.3"/>

Answer 2

[Arnaud Ceol]

You may be interested by those slides from a really good tutorial given recently during the Nettab workshop in Venice about "Semantic Web for Life Sciences": http://www.slideshare.net/sergentpepper/sem-webtutorialnettab, it provides both theoretical and practicle overview on web semantic, including RDF.

Answer 3

[DG]

RDF is basically a set of specifications that describe a dat model for metadata. The whole purpose of the semantic web is to allow information and resources on the web to be more easily and naturally processed by machines, instead of by people. RDF is just one method of doing that, but there are others. By using RDF they are essentially more easily opening up their data to any other resource that also implements semantic web specifications. The benefit is most immediately to others, not yourself.

[GreenDiamond]

Dan, I really appreciate your answer! But I still do not understand what this all means this context, i.e. in the framework of their community-based genome annotation. What data is being more easily processed by machines instead of people here? What benefits are there to making their data more easily open to any other resources that also implements semantic web specifications, in this case? I am not a computer scientist by training, and just do not understand what is being automated for this. Finally, would such an effort require starting again from scratch or adding to their current system? I have read that RDF exists in multiple programming languages? They used Ruby on Rails, BioRuby, SQL database, etc., so how can RDF be fit into that current system?

[DG]

Pierre gave some good examples of some use cases for RDF data. The simplest thing I can think of, because I am not a computer scientist either, and don't program web applications generally, is that the web as it currently exists is designed to be human readable, not machine readable. It is very difficult for computers to interpret HTML-marked up webpages, which is what is currently served up for most of the web.

The semantic web uses various technologies so that a computer (or program) can actually understand the text. Yes, RDF implementations exist in many languages, that is so you can write applications in your language of choice that can take advantage of RDF structured data and resources.

Answer 4

[krisjvr]

@Dan You replied .."RDF is just one method of doing that, but there are others". What are the other alternatives to RDF?

[DG]

OWL (Web Ontology Language) is another example. Even "plain" ol' XML can be used in semantic web applications.

[Pierre Lindenbaum]

ne4j : http://www.neo4j.org/ , ....

[DG]

I haven't used Neo4J yet but I've heard interesting things about it.