Eric Jain

Possible, maybe - but limited by the willingness of hospitals to share. - Jere

These guys are all about using aggregating publicly available data sets to achieve better results (or new discoveries) than any one of them alone could. Pretty cool stuff. Disclaimer: they're in my department ;-) - Shirley Wu

nice initial sample size: "Quantitative clinical laboratory data, consisting of 1,104,316 measurements across 656 distinct lab tests. In total, this data represented 4,844 patients across all ages that were diagnosed with one or more of a set of 12 chronic diseases... - Attila Csordas

Someone from Nature will be there and blogging the conference. Not sure who yet - most likely one of our science journalists - but watch this space so you can meet up if you like: http://blogs.nature.com/news/b.... - Maxine

Seem to have missed them. The meeting had a bit of a disjointed feel this year - or maybe it's just that it's the 6th week travelling for me. I think I'd met Magdalena a few years ago at CSHL/Genomes - she might even have still been with NRG? - Chris Cotsapas

Yes, she hasn't been with Nature for very long. She was briefly a publisher for NPG after leaving NRG, but now is back on the editing beat, I'm glad to say - she has replaced Chris Gunter at Nature. - Maxine

"A new study shows that the cancer drugs imatinib (also known as Gleevac by Novartis) and sunitinib (Sutent, made by Pfizer) halt diabetes in mice." - Maureen via Bookmarklet

It seems more like prevention to be honest - a stay of execution for the remaining beta cells if the autoimmune problem can be held back, but I'm not sure what it can do once most of the beta cells are destroyed. Of course it might open the way for replenishment (transplantation or recovery of beta cells from precursors)... not sure. Interesting though! (Disclaimer: Just my ideas, not necessarily that of my employer's) :-) - Jo Brodie

Yes, the claim for "halting diabetes" will have to be examined. It'll be good to see the actual PNAS article...this was a press splash before publication. - Maureen

I have too many entries and need to declare feed reader bankruptcy. I'd like a better way to get my journal subscriptions so I don't have to do this again. - Chris Lasher

I have few RSS feeds based on PubMed search queries - this goes as "high-importance" stuff, read from GReader daily. Also get a weekly digest from F1000 (configured around my research interests). - Yaroslav Nikolaev

I think collecting a ton of feeds over at Google reader (can take them!), and then doing a keyword search to find what you looking for works best.... Also a tag to get the "daily" feeds that you always read among the mass is good.. - Ntino

Anybody interested in setting up an microblogging server ? identi.ca above is open source and free for anyone to download and install, plus other instances "talk" to the central server... just so we can avoid central authorities :-) - Ntino

Federal recreation, camping and tour reservation information. - Eric Jain

context: http://friendfeed.com/e/8c7b43... - Michael Kuhn

(though I still don't fully understand what the group will be used for... as email directory?) - Michael Kuhn

Until we find a better solution, an email directory. - Paulo Nuin

at least 2: calculation is more transparent, and you don't have to explain what the median is ;-) More seriously, if you assume your data is normally distributed without outliers, then there is no reason to use the median. And if you return the median instead of the mean, then you should not use the standard deviation to describe the spread of the distribution but the MAD or the IQR. Hope this helps! - Yann Abraham

@eric jain nice one ;-) @dekay your flip coin example relates to binomial distribution isn't it? so it approximates to a normal distrubtion after 5 to 10 flips and then mean should be more appropriate than median ;-) thank you wikipedia http://tinyurl.com/fwn4j ! - Yann Abraham

Nice Eric. - Daniel Jurczak

http://bioinformatics.oxfordjo... or http://www.ploscompbiol.org/ar... or http://www.sciencedirect.com/s... although.. I do seem to recall a database specific (not 'resource') paper, but I can't find it either.. - Daniel Swan

Wow, Daniel, thanks! I was thinking of the Veretnik paper, which, as you pointed out, is more about services. - Chris Lasher

http://www.nature.com/nature/j... - Pierre

because a log2 ratio of 1 is equivalent to a fold change of 2, a log2 ratio of 2 is equivalent to a fold change of 4, log2 ratio of zero is 'no fold change'. Convenient eh? :) - Daniel Swan

Are factors of 2 really relevant in terms of cDNA? - dekay

@dekay: 2 fold changes in mRNA levels have largely been considered to be 'significant' changes in array experiments which are of insufficient replication to have a statistically significant result drawn from them. This is, I admit, largely historical. - Daniel Swan

Haha! This wins my award for favorite article title of the year. Particularly apropos to Deepak's talk, on multiple levels. - Chris Lasher via Bookmarklet

BioMoby users: http://biomoby.open-bio.org/in... - Jere

Interestingly neither of the two papers talk about discovering (or even reusing) existing BioMoby services. Are there any stats on the amount of reuse (e.g. how many services are used as part of how many workflows from different organizations)? - Eric Jain

There aren't enough expletives in the English language to express how deeply I agree with this. - Paul Davis

Check out Law's laws (http://bioinformatics.roslin.a...): the classic rant about this. - Jan Aerts

Can I 'like' this twice? - Simon Cockell

love the group setup: "Here, we present preliminary neurocognitive data from matched groups of entrepreneurs and managerial controls, which highlight the potential of this approach to examine advantageous risk-taking." ...entrepreneurs behaved in a significantly riskier way, betting a greater percentage of their accrued points (63%) than their managerial counterparts (51%) - Attila Csordas via Bookmarklet

"Using single-dose psychostimulants to manipulate dopamine levels, we have seen modulation of risky decision-making on this task. Therefore, it might be possible to enhance entrepreneurship pharmacologically." http://is.gd/7QMe - Attila Csordas

If any of you are in the area of Virginia Tech (Blacksburg, VA), do come out. Let me know if you'd like help making arrangements. - Chris Lasher

This inspired me as well, thank you: http://www.enroweb.com/blogsci... - Enro

Also makes a good "one figure" at www.epernicus.com, if you're stuck for ideas - Neil Saunders

I do that all the time. I can highly recommend it. Add some tagging (or branches) for versions you send around for review, and you're all set. Beats version tracking in Word. - Egon Willighagen

actually the idea of using Zotero for collaborative paper writing seems very interesting, however currently (v1.5) it only fosters single-user synchronization mode, and hacking for group sharing appears cumbersome...'d have to wait for Zotero 2.0.. - Yaroslav Nikolaev

and rethinking the above discussion (including binary vs ascii part) - it appears that would Zotero or Mendeley extend their desktop clients with a "2.0" text-editing functionality (multi-user + version control), they might actually win over the conventional workflow [Word/OpenOffice <> plugin <> reference repository]. Unless Neil will mashup Zotero with GDocs before that ;-) - Yaroslav Nikolaev

Dunno for London but perhaps http://www.science-groove.org/... may be of help. - Daniel Mietchen

Thanks for that. Though I suppose I don't really want to limit myself to those that make music about science per se, rather those who have science types in the line up. Could be nice way of organising informal discussion evening with some entertainment generated form within the community. - Neal

Various Nature editors based in London (and elsewhere) are involved in rock bands - try Henry Gee at h.gee@nature.com, or look around Nature Network (use "music"-related tag searches) - lots of related conversation, scientists involved in music, events (there is a London-based events forum), etc, there. http://network.nature.com. - Maxine

More "classical" than anything else myself, and also not in London, but could be interested in occasional things. - Cameron Neylon

"Thousands of scientists could be unwittingly ruining their experiments merely by using standard plastic lab equipment, according to a study published in Science. Andrew Holt of the University of Alberta in Edmonton, Canada, found that a disinfectant and a lubricant used in the manufacture of plastic pipette tips, tubes and micro-plates could leach into common solvents, including water (G. R. McDonald et al. Science 322, 917; 2008). Holt and his team determined that the contaminants are potent enzyme inhibitors, and skewed their drug-activity assays" - Attila Csordas via Bookmarklet

The lack of testing in bioinf software is a bugbear of mine. I've made (very small) mistakes in a published paper due to rounding errors in a scoring algorithm -- my bad. Who knows what else is wrong out there though. - Andrew Clegg

This is not new, though, was an issue when I was doing experiments. I remember a years-old paper by Alec Bangham and colleagues reporting on contaminated (distillied/deionized) water in a range of labs they visited. At that time, his mantra was that labs are simply unaware of the levels of contamination that exist. I am sure he was not alone and that others have been reporting similar issues in the interevening years. - Maxine

I amended the description with those terms, but also pointed out that all life science topics are welcome. Room managers, feel free to amend description if you can do better. - Neil Saunders

Get money from the NIH to teach people how to do bioinformatics research. - Chris Lasher via Bookmarklet

Good question! Perhaps a better measure of divergence might be some way to quantify chromosomal rearrangements, since I think those sorts of larger-scale changes (deletions, crossovers, aneuploidy, etc) might be more common than "ORF-by-ORF" evolutionary changes. - Bill Hooker

"The Yu group have created a ‘home-made’ micro-array on a music CD that just needs an ordinary CD player to scan the disk for the results of the binding assay. They did what!?! You heard right – and their system is impressively flexible since interactions involving DNA and protein can both be analysed." - Maxine via Bookmarklet

"The disk can be inserted into a standard CD player to analyse the results! Because music CDs have been designed to play back smoothly even when they get scratched or dirty, the music is encoded in a way that allows error detection and correction. The software analyzing the readout from the disk knows precisely where the errors on the disk lie. Now, normally you don’t care about the errors; you’re only interested in hearing a faithful rendition the toe-tapping tunes of your favourite combo (‘Hold on Tight’ by The Electric Light Orchestra anyone?), but with freely-available software tools, you can access the error information and that is exactly what the Yu group have done." - Maxine

Whoops! Bumped return by accident. What I meant to say was, I'm new to friendfeed and wanted to introduce myself. I am a scientific training officer at the European Bioinformatics Institute and am looking to keep up to date with developments in the field. Look forward to reading and posting in the future. - James Watson

Welcome! I haven't found a better, more fun way to stay up to date, and to discuss science matters than this room.Hope you find it just as enjoyable! - Chris Lasher

Welcome. There is also science online, a room dedicated to science communication, here: http://friendfeed.com/rooms/sc..., please feel free to join. - Maxine

ethics for sure - sofarsoshawn

Information consumable by software. Exhibit A is every paper ever written on data mining of literature abstracts. - Paul Davis

@Andrew - your term of ONS ecosystem is a fairly good representation of how we're evolving. To truly be open ONS can never be reviewed as quickly as it is published but we're getting better with multiple judges commenting on the notebooks now - Jean-Claude Bradley