Agreed with Jason. I like RDF and have played with it out of interest, but not run into lots of available data in my day to day work.
- Brad Chapman
Fair enough-- I've been sitting on gigabytes of RDF for awhile, and using it in my commercial business (just because it's the early days); I'd be happy to post a few (possibly) interesting bits to you guys to try out. My only request is that we openly work together on it to do some demos possibly. Cheer!
- Eric Neumann
To Eric and everyone who has "liked" this, please explain what this means. I understand the concept of RDF. I do not understand what you want Bio* projects to do with it.
- Chris Lasher
Chris. If the public dara were formatted/normalized using RDF, most parsers, SQL queries used to cross the data between two or more datavases would be useless. bio2rdf is a good example: http://bio2rdf.org
- Pierre Lindenbaum
Pierre, I interpret your comment as meaning "using RDF, there would be no need to develop nor maintain data and query parsers", which I pretty much agree with. I would hope that would free most informaticists to spend more time writing analytics and views, and less on bit swapping- IMHO...
- Eric Neumann
@Eric , yes that's what I tried to say with my poor English skills ;-)
- Pierre Lindenbaum
I "like" most mentions of RDF + bioinformatics, it tends to get lively :-) I interpret the post to mean "there is more out there than we realise". If so, as the first 2 comments said, let's see it! I hear a lot of talk, I don't see much in the way of usable data or tools.
- Neil Saunders
++ to jason, brad and neil's comments. bio2rdf is a good start, but why do we even need RDF? BioMart is another warehouse that seems to work nicely on good old SQL. I think when people see a tool of comparable utility that depends on RDF they will take notice.
- Chris Mungall
Chris- I actually wasn't referring to bio2rdf- as you said it's a nice start, but not key for most imformaticists; alternatively, I have converted GEO, mutdb, and pathway data into rdf. Was hoping you guys would find this useful to try some investigate?
- Eric Neumann
from iPhone
I'd be interested in checking it out - but I've already paid the RDF tax. What the average bioinformatics hacker would benefit from would be a paper showing off some of this stuff with some clear non-waffly reasons why rdf works better (or not better) than sql, json, plain xml, couchdb, etc.
- Chris Mungall
A demo for RDF I'd like to see: extracting a list of genes from a RDF datastore for a given organism in a given region (chr:start-end). demo n°2: thoses genes must be the descendants of a GO term.
- Pierre Lindenbaum
"RDF tax" is a funny expression; do you mean RDF-XML or RDF in general? Most of my projects don't bother with RDF-XML; they're either RDF-N3 or RDF-JSON... remember, RDF is not a syntax, it's a data source and type binding semantic! As I said initially, I think most folks have missed what RDF is really capable of (and where it saves time). I will post my data shortly...
- Eric Neumann
I think we miss what it's capable of because we don't see demonstrations. My impression with many data formats, ontologies etc. is that once designed, their creators say "OK, we've made you a nice format, you all have XML/JSON/whatever parsers, get to work!" So here's my plea: when you create the format, create a small tool that demonstrates its worth, make it public and advertise it.
- Neil Saunders
Should direct this here in this thread also: How many of you know about/use MIT's Exhibit?
- Eric Neumann
Now, that's what I'm talking about. I didn't know about/use it before - but now I do/will (or might). Thanks!
- Neil Saunders
I'd be more than happy to share all my examples and code extensions of it to all of you... some simple examples under "demo" at http://eneumann.org/
- Eric Neumann
btw, just look at the source HTML of these examples, and that's all there is to creating your own facet RDF viewing app ("steal this code")
- Eric Neumann
OK, spent a little time playing with Exhibit and I am now officially really impressed. Giving me lots of ideas. But I'm wondering how large a JSON file can be? Hundreds, thousands, millions of items?
- Neil Saunders
Thank you to everyone who helped and encouraged on this and offered future support! Among many things, one thing I'm a bit embarrassed about is the lack of details in the "broader impacts" and open science parts of the proposal. Definitely it needs more planning and better writing, and the panelists are going to point that out. Hopefully you all understand that it's a product of spending all my time trying to make the research parts as good as I could.
- Steve Koch
Thanks everyone! I sort of dropped off the map after submitting the proposal. One of my most successful attempts at actually vacating during a vacation.
- Steve Koch
Bummer! Just found out tonight that the proposal was not funded. I have done one read-through of the reviews. First of all, I am once again amazed at the time and effort the reviewers put into reviewing my proposal. There were 6 reviewers in total and they all had a page or more of feedback. And they all pretty much agreed on the main points: (a) interesting proposal that is worth...
more...
- Steve Koch
Thank you Jean-Claude, Andy, Cameron, and Drew for the Open Science support letters! I am sure that was key to the very positive reviews for Open Science. Thank you to everyone else on friendfeeed who supports me and our other lab members. We now have 7 months to obtain the preliminary data and get a couple publications, and indeed we are now poised to do so. I am confident the lab members can do this, and by eliminating (b) and (c) above, all signs point towards this being a very strong proposal next time.
- Steve Koch
well done all 'round! next time you'll knock 'em dead.
- D0r0th34
for the CAREER third time is often the charm - good luck for next round
- Jean-Claude Bradley
Thanks, Jean-Claude and Dorothea. The CAREER "3rd time's a charm" effect is part of the reason I submitted without strong preliminary data. (But I also thought we'd be able to generate the data in time for the 2nd submission. Now, though, I'm SURE we can before July :) )
- Steve Koch
After thinking about it for a couple days now, I am still very happy that they viewed the open science so positively. I think that's a big deal.
- Steve Koch
absolutely Steve - very good feedback about OS being a positive to the funders
- Jean-Claude Bradley
Steve, this looks really really promising. If you can come up with interesting data for the next resubmit, I think you have a great chance to get funded. Congratulations!
- Bill Hooker
If I have my story right I think this came out of a criticism from a review panel that the structures and computational bio department was not collaborating enough. They came up with the mycoplasma collaboration that Luis Serrano in particular was very excited about. 3 science papers is not a bad way to show results :). I still have to read them.
- Pedro Beltrao
Mycologia, Vol. 101, No. 6. (c 2009), pp. 751-763. Here we report the characterization of fungi from 10 accretion ice sections (3300-5100 y old) as well as two deep glacial ice sections that are close to the bottom of the glacier (1,000,000-2,000,000 y old) from the Vostok, Antarctica, 5G ice core. Fungi were characterized by fluorescence microscopy culturing and sequence analyses of ribosomal DNA internal transcribed spacers. A total of 270 fungal colonies were cultured from the accretion ice of subglacial Lake Vostok and an additional 14 from the glacial ice immediately above the accretion ice. Mean concentrations were 0-4.42 cells mL(-1) ice meltwater of which 0-100% exhibited viability (as determined by fluorescence microscopy). Thirty-one unique fungal ribosomal DNA sequences (28 from accretion ice and three from glacial ice) were determined and compared to recent taxa. The results, plus tests for growth at low temperatures, indicated that Lake Vostok contains a mixture of...
- Neil Saunders
ISMB 2010 is accepting high-quality research papers in any aspect of computational biology. Accepted papers will be published as conference proceedings in an open access, online-only, section of a regular issue of the journal Bioinformatics.
- Adam Kraut
Posts with working code examples rock: * Using Elastic MapReduce to access SimpleDB on Amazon Web Services. This uses the boto Python library to access the database and a Ruby client, configured with JSON to run the example on Hadoop: http://atbrox.com/2009... ...
- Brad Chapman
from Posterous
There have been a number of great posts recently with visualization tips: - Rajarshi describes customizing R heatmaps, which are immensely useful for things like visualizing next-generation sequencing results across tiles: http://blog.rguha.net/?p=419 - Revolution computing reviews R code for ti ...
- Brad Chapman
from Posterous
the time series visualization is pretty cool
- Rajarshi Guha
It's a lot of work. Probably easiest to have a clear landing page that says "here are all the other places to find me".
- Neil Saunders
Yes, more or less figured that out. Question of how to set that page up at the moment. I do want to aggregate some stuff back to one place though. OpenIDs, FOAF, draft documents, that kind of thing, and some view of the streams that I'm generating. Aside from anything else I need to learn how to actually run/deploy/setup these things.
- Cameron Neylon
On a related note: It seems you used to be able to generate FOAF for your Friendfeed subscriptions using http://friendfeed.com/<username>/subscriptions?output=foaf .. but unfortunately it doesn't seem to work now.
- Andrew Perry
A domain name is just a name so if you've got some content hosted elsewhere already you can just use it as an alias - however I think the problem you're actually talking about is hosting that content. Assuming you're not already tied into one particular hosting company (as part of the domain name registration) I'd advise looking at a few content management systems (although they're...
more...
- Dan Hagon
Wait, friendfeed FOAF doesn't work anymore? Data portability fail there! I hope that's not a facebook-mandated change.
- Mr. Gunn
Cameron, I'd look into delegating your domain URL as an OpenID URL. Verisign has a pretty cool profiles service: http://williamgunn.pip.verisignlabs.com/ and there's also Google Profiles.
- Mr. Gunn
@Mr. Gunn - fancy stuff, this, very cool. Coverflow-style scrolling borrowed from OS X. Methinks I should switch from MyOpenID (http://mummi.myopenid.com)...
- 'Mummi' Thorisson
Yep, openID is coming. Think I'll be doing something quite simple with a Wordpress install as a starting point.
- Cameron Neylon
Heya! Do any of you know where to find one of those graphs that show how much sequence data has been deposited in sequence databases (preferably recent ones)? Like this one: http://farm1.static.flickr.com/15... , but more recent.. I would be much obliged!
And the liquid ban is BS anyway, for several reasons, including: (1) confiscated liquids are disposed of as if they are completely harmless; (2) there are no consequences for attempting to bring liquids on [see also: http://www.schneier.com/blog...]
- LogEx
sorry, I laughed when I read your tweet :-) How big are the data now ?
- Pierre Lindenbaum
Big enough :-) It's actually to let our students at a programming course do some annotation on SNPs, and it would be nice when they would get the data just from the 1000genomes website. I know, it's sad :-)
- Jan Aerts
from email