Sutee Dee › Likes

Chris Lasher
What are you using to visualize biological networks? -
What are you using to visualize biological networks?
[Image: "network" by Simon Cockell, link to original source for attribution] In particular, what are you using to view networks in a dynamic way? Cytoscape seems the behemoth for visualizing biological networks, but have you used other solutions? A group mate and I have been discussing how we could *really* use a JavaScript library that rendered as SVG in a web browser. This would allow us to call for more data from our databases and present data on the fly and only-as-needed. He's convinced we could start writing one; I'm concerned it's a lot more difficult than he thinks, but we're both fishing for other ideas. - Chris Lasher from Bookmarklet
for working with SVG and Javascript, maybe try svgweb or Raphaël - Mike Chelen
We're using Raphael in the Protein Geometry Database <> for rendering our graphs. It's been a huge boon to cross-browser compatibility since IE *still* can't do SVG. - Donnie Berkholz
You might try a particle demo on Chromium's canvas element. Other than that I think the optimizations for performance could end up being the bigger project. Not undoable, just maybe not worth the time. Also, specifically Chromium as they're canvas element is current the best, though hopefully FF and Safari aren't too far behind. For IE, just give them a link to Chrome Frame ;) - Paul J. Davis
Is this meant for big graphs? Is the focus on layouts etc? Prefuse is a pretty nice toolkit for graph layout/calculations. igraph also is an excellent library with a bunch of layout algos - Rajarshi Guha
Many thanks for your input, guys. @Rajarshi We have graphs of 5k-12k nodes and 20k-80k edges. I don't know if that counts as large but they're not small. @Paul I'm worried about performance (responsiveness) given that Firefox can take a pounding in some of the demos at That's why I wonder if Cytoscape is really the only option (right now). @Mike and... more... - Chris Lasher
Thanks, Chris! I certainly hope it enables some folks to do interesting science instead of just looking pretty, though. =) - Donnie Berkholz
Jan Aerts
Thinking of unconventional ways to visualize exome resequencing results to search for causative SNPs using #processing
Don't know yet if I can come up with one/some... But trying to keep an open mind. - Jan Aerts
That's what I'm missing: a good idea for an unconventional way to visualize a *omic thing :-) - Pierre Lindenbaum
This comes back to the discussion long time ago if modern big data research needs graphical designers/artists to help visualize the stuff. We might have the data, but not the skill to visualize (other than in histograms/piecharts/DAGs). - Jan Aerts
I prefer it if a complex dataset can be distilled down to a simple bar or scatter plot. I find that too much data can distract from the main point. I bought and read the Visualizing Data book but I wasn't moved to try and start creating things in processing. - Michael Barton
Having said that though I find the R syntax very frustrating. I would love a combination of R functionality with Ruby readability and metaprogramming. - Michael Barton
That visualization of a set of SNPs was great & original - Pierre Lindenbaum
Although I also like simple statistics, I feel that showing everything to the user can be really useful because it can give a better "sense" for what's going on. But it's difficult: the data must be presented in such a way that trends present themselves, rather the researcher having to define the trends and then drawing them. The data must draw the patterns to the screen; not you. (If you know what I mean....) - Jan Aerts
Khader Shameer
Need to write about TP, FP, FP, FN, Sensitivity, Specificity, Accuracy, Positive Prediction Value, Negative Prediction Value in my thesis - would like to know if there is any authentic references available for them or just to cite Wikipedia or Mathworld ?
This list should give you a start: - pn
Thanks Paulo, am going to check those papers. - Khader Shameer
Ryan - @magicofpi
Windows 7 Upgrades Chart Makes Brain Hurt; We Offer the Mac Equivalent -
Windows 7 Upgrades Chart Makes Brain Hurt; We Offer the Mac Equivalent
Windows 7 Upgrades Chart Makes Brain Hurt; We Offer the Mac Equivalent
Show all
又一个m$ VS mac,又一次mess VS simple - iJay from email
Andrew Chen
When my macbook is hooked up to the LCD TV, this Twitter status box is like 2 feet x 6 inches. Megatweet!
Andrew Clegg
Just convinced my boss that it's a much better idea to let us manage the group 'virtual library' in Cite-U-Like, rather than by copying-and-pasting citations into emails to our long-suffering curator (who I think is the most pleased out of any of us)...
Congrats! Lets hope for your sake he does not talk to someone and changes his mind again :) - Dilip Dand
She. But I don't think she will when all the nice extra features become obvious... - Andrew Clegg
Andrew, can I suggest you spend a group meeting working with people to show them how to use citeulike effectively? It really helps if everyone is on the same page with understanding how people think about what they are doing (and know how to book mark stuff). Important with these things that you get some quick wins and a rapid start off - otherwise some people will tend to revert to the old approach and then it gets messy. - Cameron Neylon
Yeah, good idea... - Andrew Clegg
Does anyone have any experience with the 'tag selected articles' feature? It doesn't seem to work properly -- just adds the tags to the articles in your own library, even if you do it from a group library - Andrew Clegg
Just "fixed" the group tagging thing in CiteULike - Fergus Gallagher
Iddo Friedberg
Deepak Singh
Episode 5 : Grand challenges, javascript frameworks and collaborative annotations -
Any plans to revist javasript frameworks (now SproutCore is 1.0, MobileMe might be getting a revamp, progression with Objective J etc)? Or thoughts on a pubmed podcast? - Tom Tubbs
Probably will. Perhaps in the context of HTML5 as well - Deepak Singh
Attila Csordas
Do you know a practical, intuitive, easy-to-learn formalism to uniformly describe molecular regulatory mechanisms such as induction/activation/inhibition/suppression et all, pathways & feedback loops? A formalism that can serve as a pseudocode for a code playing w/ these molecules & pathways? Here's a paragraph from a recent p53 progress paper,... "Multiple mechanisms have been invoked to explain how p53 triggers MOMP. MOMP is usually inhibited by anti-apoptotic multidomain proteins of the Bcl2 family (such as Bcl2, Bcl-XL (also known as Bcl2l1) and Mcl1), and is conditional on pro-apoptotic multidomain proteins from the same family (in particular Bax and Bak (Bak1)) that can homo-oligomerize within the outer mitochondrial membrane to form MOMP-mediating supramolecular structures. Depending on their particular affinities for multidomain Bcl2 family proteins, a set of distinct pro-apoptotic 'BH3-only' proteins can directly interact with Bax or Bak to trigger their homo-oligomerization and hence MOMP (these BH3-only proteins are referred to as 'direct activators') and/or neutralize one or more anti-apoptotic multidomain proteins (referred to as 'sensitizers' or 'de-repressors'). In contrast to direct activators (in particularly the proteins Bim (Bcl2l11) and... more... - Attila Csordas
can anyone extract it into a 2-3 lines of formalism for instance? - Attila Csordas
Isn't that what SBML is for? See for example here: Of course I may be wrong - for example I don't understand the difference between SBML and BioPax (another format in which BioModels database stores its data). - Pawel Szczesny
SBML hosts a matrix of software you can use to build (among other things) your own models: - Pawel Szczesny
BioCyc is worth looking at for real life examples of formalized pathways ( They have a open source relational database schema in BioWarehouse (, in addition to SBML and BioPAX formatted files that Pawel mentioned. - Brad Chapman
What's weird is that I was just last night thinking that biology is not going to make any more big advances until we have a (computable, abstract) formal logic for molecular pathways. Then I started wondering about similar formalisms for, say, physiological processes like homeostasis, and ecological processes, and so on. I knew about SBML, hadn't heard of BioPax/Cyc. - Bill Hooker
Which is to say, Attila, that I don't think the thing you are looking for really exists yet. - Bill Hooker
thx for the smart tips, Pawel & Brad, @Bill good to see that our ideas converge, every time I read a paper on pathways I think that the linear text is just plain wrong in explaining the topic and redundant and a proper & significantly shorter formalism is needed. - Attila Csordas
Anyone who's ever tried to get to grips with a pathway has had to invent their own representational logic (arrows, circles, etc) -- there's just no other way to map this stuff out and get your head around it. It's a short step from there to realise that an abstract formalism is necessary for logical manipulation of pathway information -- without the formal structure, you can't keep... more... - Bill Hooker
You should try SBGN, its best way to formalize the pathways or any network activity, it provides 3 different language - Process diagram is one of them. SBGN is new and I guess it will be great if every biologist learn SBGN. - Abhishek Tiwari
You could check out Kappa, the formalism for signaling logic that Walter Fontana et al have developed and that Cellucidate is commercializing. See PNAS paper linked at I've no idea if it's easy to learn or not. - Mikael Huss
the challenge is still open to formalize the p53 paragraph above - Attila Csordas
Jeremy Zawodny
JavaScript syntax fucking drives me bonkers.
The prototype stuff never quite takes for me. - Todd Hoff
Wait, the syntax? What's the problem there? The syntax seems pretty conservative and straightforward to me. Object semantics are a bit unconventional, the DOM is hell, but the syntax? - ⓞnor
That doesn't seem like a syntax complaint to me. But I guess the syntax you use in JS to construct something like an ordinary class ends up being pretty tortured, mostly because, well, JS doesn't actually have anything like ordinary classes. - ⓞnor
Isn't that what "elif" is for? - ⓞnor
How about this little gem? function foo() { return [pretend there's a CR here] "bar"; } alert(foo()); => "undefined" - Joel Webber
Ryan Block
Top complaint about Kindle: too expensive. Amazon's solution: make an even larger, more expensive model for broke students.
Academic books are $60 to $120 each and weigh many pounds. If they can get even half of the academic books ported to the Kindle this is a big win for students, even though it is expensive. - Robert Scoble
I can imagine universities giving Kindles to first-year students or at least offering them to students at a reduced cost. I don't think we'll see a huge reduction in the price of text books, because you will still be paying for publisher costs. - Nathan Finley
My books from school had a lot of text marked in yellow or pink, to remember what was important. I think it would take a lot to get use to using the Kindle - Asgeir
Asgeir - you can bookmark, highlight and annotate text in Kindle - Jamie
scoble: are the text books free then? - Riaz Kanani
@Asgeir It is really only a matter of a couple of years that students will have NOT known that Kindles and other readers haven't always been around. To them it will be a natural to tag, search, share, etc. - Jim Espinoza
It really is expensive. And you know textbooks are still going to cost $100+ each. - Barry Biddlecomb from twhirl
If the books are cut down to about half the price because of the kindle, every student will get one. I need to spend 200+ on books every quarter, and that's with buying used and looking for the best deal. - robbie nakamura
the text books don't have to be *free*. if you spend $900 on textbooks a year (, the Amazon textbooks just have to be 14% cheaper than the paper versions (saving you $126/year) for the Kindle DX to pay for itself after 4 years. - Karim
good analysis here of why it's good for the publishers, too: basically, without printing and shipping costs, they can sell books at lower cost but with a higher margin. Amazon is trying to do to books what Apple did to music. :-) - Karim
good point karim - ebooks are cheaper typically than books.. so could be good if text books see the same level of discounting - Riaz Kanani
so what is going to happened to libraries, if books in the future is digital?. Perhaps one day we can rent a book and read it on the Kindle? - Asgeir
The Kindle is still a novelty until it comes with colour and supports graphic novels/comics, with backing from Marvel and DC. And drops to about $199. Then they'll be flying off the shelves. As it is now it's an expensive gimmick. What amazes me about the DX is you now need two hands to read. Can't be long until the market is flooded with cheaper, smaller alternatives. Or more likely... more... - Shéa Bennett
Kindle should come in 2 flavors IMHO... Wireless G and EVDO. I already have sprint internet on my phone. I won't buy a kindle to pay for a connection I already have. - Gunneh-MoTo Calls me that
Students already have laptops: they need PDFs not another large device to lug. Physical books sell used @ 50% eBooks are new only - Tyler Ham
My experience in the UK at least is that the ebook versions are not cheaper, can usually find printed ones heavily discounted. Although we can't actually get the Kindle, only other brands like Sony and have to buy ebooks from Waterstones or WHSmith. Will Amazon launch in UK soon? - Steven Horner from Nambu
I'd love something that is, well, like a book. Small and foldable. I couldn't take this on the train to read. People are used to reading books why not design to feel like one. ;-) - Kol Tregaskes
Amazon is still only aiming this at the American market, once they go global, it will be quite a different story. - chris demeyere
Is 500$ too much for being able to sit in the sun, save trees and finally read programming papers without code snippets getting messed up? - Thomas Amberg
If my textbooks were available on a kindle... I'd pay up to a grand for one. The kindle is reusable from year to year... my text books... not so much. I pay just about two grand a year for text books. - Joshua Schnell
Core texts are significantly cheaper in the UK, and unless you can resell on your core texts second hand (which you don't seem to be able to do with an electronic text in this case), it's going to be a difficult case to prove for the academic text in Kindle - David Bird
you can resell a second-hand Kindle just like a physical book. :-) i'd guess that licensing on the DRM'ed books doesn't allow resale, but i can imagine a black market for those. how much would a freshman pay for a 4-year-old Kindle that already had every single textbook they were going to need in the next four years on it? :-) and what if the textbooks were already annotated with "this question will appear on the final" notes? NOW how much would you pay? lol - Karim
Karim, that's genius. Could you imagine universities and colleges building them into their programs? Students would get a preloaded kindle with their course materials in 1st year. Man, that would be a game changer. - Joshua Schnell
It's a no-brainer for all economic reasons laid out by many above commenters as replacement for dead tree textbooks. So only shoe to fall is MIT, UC Berkeley, CalTech and Princeton (above all- Bezos alma mater) to be part of a discount university marketing program launched by Amazon. Bring it on. - Lisa thorell
Attila Csordas
A suggestion: FriendFeed type of Q&A (uninterview) with cool life scientists (non-celebs, celebs) here using the initiative & weight of this community? Algorithm: i., picking names ii., set up a timeline iii., sending out invites iv., posting a message (body of the Q&A) v., discussion 1 example is how I invited Steven Brenner to contribute... by now we probably have the strength to reach & invite most of the people we'd like - Attila Csordas
Good idea, Attila! I agree that FF would be a suitable platform for iii. For i and ii, I think a blog or wiki would do better. What about ? - Daniel Mietchen
Daniel, the idea is to have an uninterview w/ a life scientist here (every step included) harnessing the energy of this community. (iii., sending out invites means inviting the uninterviewee to FF in order to do the Q&A here), testing FF as a platform for such activities - Attila Csordas
That's actually a pretty good idea, Attila! - Mr. Gunn
I think you'd have to give your interviewee a reasonable timespan -- say, two or three days? -- during which he/she agrees to check the FF thread and post replies as often as time permits. Otherwise you might as well be doing it by skype/irc; I think the advantage of FF is the small but useful time lag, which allows time to think about answers and time to fit the uninterview into a busy schedule. - Bill Hooker
Bill: you're right, we have to agree on the proper timespan of the Q&A. I think 2 or 3 days sounds perfect. We should ask interviewees to allocate, say 1 hour time/day to engage in the discussion here & answer at least, say 3 questions/day. How does that sound? Maybe a bit overregulated? - Attila Csordas
Thanks, Mr. Gunn, hope others will like it too. - Attila Csordas
Attila: maybe a bit over-regulated; I figure if we ask interesting questions the interviewee will end up spending way too much time here, like we all do. :-) - Bill Hooker
Let's pick some scientists (academia, industry), who would you like to (un)interview here? - Attila Csordas
Atul Butte, Paulien Hogeweg, Genentech's Art Levinsohn, Craig Venter (reached his PR agent once) - Attila Csordas
that's awesome! great idea! - Alexey
Declan Butler (Nature/France), Jean Weissenbach (CNS/France), Catherine Letondal (Pasteur/France), Pierre Tambourin (Genopole/France), someone at NCBI, other fields of science (humanities, physics, etc...) - Pierre Lindenbaum
Declan (a colleague) is a journo, though, not a life scientist ;-) Do you want to broaden it to hangers-on (camp-followers) or keep it to proper scientists? - Maxine
there's no need for a 'proper scientist' vs. journo distinction, everybody who's doing interesting stuff in the bio/bioinfo/biotech/biomedical field qualifies as an (un)interviewee - Attila Csordas
Thx Pierre for the suggestions, other cool biopeople you'd like to invite here for a 3 day interview period open for any kind of questions? - Attila Csordas
How about James Watson? - Jim Hardy
How about turning the tables on Richard Poynder? ( I suspect all the blogs-vs-msm people would have plenty to ask Richard. - Bill Hooker
Cosma Shalizi ( makes occasional forays into biology and would probably have lots to talk about with programmers and statisticians hereabouts. Data mining, social and other networks, machine learning, signal transduction, collective cognition and dozens of other things I don't understand: - Bill Hooker
That's another good one, Bill. Also, maybe Wolfram? He seems to be doing a bit of a PR tour ATM. - Mr. Gunn
Georgy Daley is a great speaker, and Rudy Jaenisch would be another recommendation. - Mr. Gunn
Jeremiah Owyang
Which websites could really benefit from Facebook Connect? I say: Pandora, Fandango, Yelp, to start with
Amazon,, Zappos - Jesse Stay
iTunes - Jesse Stay
Pandora + Facebook Connect would be great! Personal Algorythm for your music taste meets Social Influence - TedBailey
Ecademy Linkedin Xing Naymz - Thomas Power
Hi5 Perfectnetworker betternetworking telecomyou blip - Simon Hamer
Attila Csordas
Using Engineering Principles To Study and Manipulate Biological Systems -
Using Engineering Principles To Study and Manipulate Biological Systems
presenter is a systems biologist turned Google Search analyst surprisingly - Attila Csordas from Bookmarklet
Abhishek Tiwari
The information flow model for Protein-protein interaction networks -
The information flow model for Protein-protein interaction networks
Thomas Hawk
For Some, Voice Mail Is Losing Its Allure - -
For Some, Voice Mail Is Losing Its Allure -
For Some, Voice Mail Is Losing Its Allure -
"Tim Kassouf from Baltimore, 24, who calls himself “a certified voice mail hater,” said he had 68 messages, 62 of them unheard, in his cellphone mail box. Scott Taylor, 41, a senior manager at an e-commerce company in Phoenix, said voice mail was “just totally an ineffective communication method, almost ancient now.” Like many others, Mr. Taylor advises callers on his outgoing message to try his cellphone or to send an e-mail message if they need to reach him right away. It is good advice. Research shows that people take longer to reply to voice messages than other types of communication. Data from uReach Technologies, which operates the voice messaging systems of Verizon Wireless and other cellphone carriers, shows that over 30 percent of voice messages linger unheard for three days or longer and that more than 20 percent of people with messages in their mailboxes “rarely even dial in” to check them, said Saul Einbinder, senior vice president for marketing and business development for uReach, in an e-ma" - Thomas Hawk from Bookmarklet
I sooooo hate voicemail. It's the worst thing. I hate the nagging feeling that you ought to be checking it, long winded messages where people won't simply leave you a phone number, having to have a pen to write down a number. I wish voicemail would die. People should use email or other electronic means of communication over voicemail. Come to think of it, I pretty much hate the telephone all together. - Thomas Hawk
I solved this problem: I have a message that says not to leave me voice mail. It's interesting how few people do. - Robert Scoble
Hmm funny, I think about going voicemail-only with my phone. As in: People can never call me. - Meryn Stol
i have not checked my voicemail in over two years. the outgoing message warns people of such, yet people still leave messages, i see them there, but i am not nearly THAT curious to see what someone who clearly cant comprehend that I MEAN it when i say i don't check it has to say. FUCK'em. - merkley???
I have removed the voicemail facility on my number and replaced it with a message telling people to email me, a much better solution - Arthur Guy
at least with visual voicemail you can see who it is before deciding to skip it or not. What really pisses me off though is when people block their number. No way in hell I'm ever answering any call with a blocked number. Most likely I will also not listen to that voicemail either. - Thomas Hawk
Nice mention of PhoneTag, which I've been very happy with. I'm looking forward to try google voice's voicemail transcription service, too. - David Vasileff
I looked at our Vonage account a few days ago. We had 132 voicemails on our personal number from Dec. on. I deleted them all without listening to them. Everyone knows that to reach us you either call our mobile or email/text us. From our business # I had 75 voicemails from Dec. on and they were all the same, "I just emailed you this information" then a rereading of the email they just sent. Sigh. On my cellphone, I stopped checking my voicemail in Sept. I'm maxed out now. - Anika
I'm very slow at check home voicemails, but my mobile gets checked as soon as I see there's a message there. Think there is still a use for voicemail, but its certainly not a favourite form of messaging of mine . . . - Ed Richardson
I find SMSs a very potent alternative to voicemail. In the US, the user base that communicates via SMSs is miniscule but here in India, it is omnipresent. - Parth Awasthi
I hate getting voicemails and having to check it. - Alan Le
so do i - i hate phone calls altogether.. the expectance of immediacy by people is outrageous - especially in the cell phone age - andy brudtkuhl
Google Voice. Free. Transcribes voice mails and e-mails text to you. - Rod Bauer from twhirl
Rod - if you can get in! I've signed up a dozen times.. need a solution for my biz asap - andy brudtkuhl
I find myself not checking my voicemails because so many of them are junk. They're not technically spam, but they tend to waste my time. Visual Voicemail at least gives me some flexibility on when I want to listen to which message, but it's still an annoyance. - Kevin Elliott
Most people know not to leave me a voicemail. Their number pops up on the caller ID. I know they called and I'll call them back when I have a moment. Voice is a crappy medium for leaving any kind of information. - Eric P
Visual VM ftw. - Mona Nomura
Eric, I totally agree. (I also hate vlogs for the same reason, esp when they're just talking into the camera, as it's basically a slower and less helpful way to get one's message across.) - Andrew C (✔)
I want visual VM. How can I get it? I also want google voice. I check my cell phone voice mail on my iphone. At least it lets you see who it is from. - Amani
It costs me to listen to my voicemail on T-Mobile so if I know who it is (i.e. from the last caller number) then I don't bother to listen to it. I also don't tend to leave voicemail. Most people see the missed call and call you back anywhere. - Kol Tregaskes
The pattern that invariably happens is that somebody will call me in a meeting. I won't be able to answer it, because that's just rude. Then I'll wonder the whole rest of the meeting what the message is. - Wirehead
Mona Nomura
Hulu tip: let the show run completely to the end (w/out watching), then go back and watch all the way through with no ad interruptions. ;)
Nice tip, Mona! I'll have to try that. - Beau Liening
I do it all the time. Hulu isn't smart enough to stop like ABC's player (where I watch Lost) - Mona Nomura from fftogo
I knew there was something you were good for Mona. **goes to look for Homework assignment** ;-) - Amani
Very nice, thanks. - Kate
To add: even if the volume is off in one tab, it can be turned up in other tab. Gonna go watch DollHouse now. Excited. :) - Mona Nomura
You have saved me some serious time Mona. Awesome. - Mike Nayyar
or you could jump to each ad first, then you dont have to wait for 30 mins or whatever the show/movie length is. - Varun Shenoy
More than time, I hate interruptions :) ++ potamus Though I prefer to let it run all the way through, which requires less effort from my end ;) - Mona Nomura from IM
I'll try that if Hulu ever comes back to Boxee - Robert Hafer
Jan Aerts
Combine data visualization and statistics to understand data; don't use just one of them. -
Chris Miller
MIT adopts a university-wide OA mandate -
From the article: "Another of the world's great research universities has an OA mandate. Unlike Harvard and Stanford, which now have OA mandates for some of their schools, the MIT mandate is university-wide" - Chris Miller
This is a good thing. - imabonehead
Chocolate Pasta with Raspberry Coulis and Grated White Chocolate -
Chocolate Pasta with Raspberry Coulis and Grated White Chocolate
I'd probably like this if it didn't replicate a savory pasta meal. My brain finds it hard to extract the two. - Derrick
Hematopoietic scheme models - why do we need that many? Hematopoiesis: Blood Stem Cell & Lineages -
Hematopoietic scheme models - why do we need that many? Hematopoiesis: Blood Stem Cell & Lineages
Hematopoietic hierarchy schemes - scientific masturbation or flow of progress? - Alexey from Bookmarklet
I'll admit I'm somewhat biased given that I worked in an offshoot of Weissman's lab (and I consider him to be the expert of all things HSC-related), but it strikes me that these models present some relatively testable hypotheses: do CMPs exist? do GMPs derive themselves from CMP? do MkEP's differentiate directly from ST-HSC's? - Benjamin Tseng
@ Benj - there are many studies confirmed existence of progenitors (CMP, CLP, MEP...), not only from Weissman group, their relationships and hierarchy is not very clear and under discussion. That's why many labs - mane protocols - many schemes. I have no doubts in Weissman's expertise. He actually accept some alternatives in the canonical scheme, showed for example in his recent review... more... - Alexey
Paul Stamatiou
Ryan's Scraps: Named Scope: It's Not Just for Conditions, Ya Know? -
Roderic Page
How FriendFeed uses MySQL to store schema-less data - Bret Taylor's blog -
After some deliberation, we decided to implement a "schema-less" storage system on top of MySQL rather than use a completely new storage system. This post attempts to describe the high-level details of the system. We are curious how other large sites have - Roderic Page
Shirley Wu
Patients' genetic profiles can help avoid incorrect dosing of a common, dangerous drug, say Stanford scientists -
International collaborative effort spearheaded by my advisors. They gathered together a consortium of research groups working towards the same goal - an effective dosing formula for warfarin based on pharmacogenetics. - Shirley Wu
"Altman, Klein, Stanford medical student Hersh Sagreiya and members of the consortium collected a variety of demographic, clinical and genetic information from patients at risk for stroke, heart disease or other blood clotting problems for whom the ideal doses of warfarin had already been established by trial and error. In particular, they determined which version of each of two genes... more... - Shirley Wu
"They combined this information with the pre-determined ideal warfarin dose in about 4,000 patients and used it to develop a computerized dose-prediction algorithm. They then tested the new algorithm with and without the genetic data on the remaining 1,000 or so patients." - Shirley Wu
The "pharmacogenetic algorithm yielded predictions that were, on average, within about 8.5 mg/week of the patient’s ideal dose; using just the demographic and clinical data predicted doses that were within about 10 mg of the ideal weekly dose." - Shirley Wu
Yes, more like this! - Mr. Gunn
And here is the actual paper: - Shirley Wu
So, what's the clinical significance of the 1.5mg/week (if I'm reading this right, that's the variance decrease)? Istm that's what the FDA will be asking before approving genetic testing for clinical use in this case. - Chris Cotsapas
The clinical significance of improved dosing estimates isn't clear yet - they conclude that the study lays the foundation for a prospective trial assessing whether getting the initial dosing estimate closer to the eventual stable dose leads to better outcome or not. They also mention that the pharmacogenetic algorithm provides the most improvement for those requiring low or high doses of warfarin and not for those in the intermediate range, compared to the usual clinical and fixed dose algorithms. - Shirley Wu
[I guess I should read the paper, but...] Does the better predictor translate to faster time-to-optimal-dose? afaik that's the crucial thing with warfarin: patients not on effective doses either clot again or bleed out, so feeling your way towards a patient-specific efficacious dose is a time-limited exercise... - Chris Cotsapas
I don't think they addressed that specifically, but I would imagine it does. It should also reduce the risk that the initial dose is life-threateningly inappropriate. - Shirley Wu
Robert Scoble
New Feedly combines Google Reader, friendfeed, Twitter in great way for social network addicts -
Here's the video I did yesterday with Edwin who shows me what Feedly is all about. Edwin talks about the real time web (Steve Gillmor alert) and demos the latest. - Robert Scoble
This is really great. I tried Feedly when it was released around the FF 3.0 release and struggled with it's magazine metaphor. I was very much used to the inbox metaphor of google reader. However, the new real-time search integration with Friendfeed and Twitter is amazing. It has really become a conversation awareness tool for browsing. - Dave Senior
Used Feedly before. Gonna have to give it a look again. The idea of everything in one place is very appealing. - Paul Wade
Not to mention the great little Feedly mini bar at the bottom of every page - easy to share, tweet, email whatever you find - even outside of Feedly. INtegration with delicious, tumblr, google reader, friendfeed - it's fantastic - James Hull
Just to add here to the love - the documentation is kind of confusing at first. There are a ton of great shortcut keystrokes that I've been able to find - 'd' while reading an article, automatically opens up the delicious tagging extension (if you have it installed) - 'gg' will open a window with quick access to your feeds, and so on - James Hull
I love feedly use it all the time especially the new features - Kim Landwehr
Cool. I got here from the Feedly mini that popped up while I was reading Scoble's blog post about this. - Tom Landini
I really want it to figure out the most popular posts for me though. The digest should show conversations like Kosmix does, based on my FF/twitter/digg/reddit friends. - Brandon Titus
gonna have to give this a try asap - andy brudtkuhl
Been having fun playing with the new Feedly mini. Not working consistently on sites, and I have yet to click a link in Twitter and have the Twitter integration kick in. Still a very nice little tool. - Sean Brady
Feedly is very good. I just wish I could use it on Safari. - Louis Gray
Sean: re: consistency. we are adding a button to manually pop up in the cases where the metadata does not pass the "interesting" bar. Regarding twitter, can you please open a bug on and let us know which twitter page you are clicking through from and if you have any other twitter extension which might interfere with the topic detection and we will track the issue down. - Edwin Khodabakchian
thank you Louis! Today reminded me of the launch you helped us detonate back on June 16th (and hope that we have learned our lessons and nothing will backfire this time!). Thank again for the initial push! - Edwin Khodabakchian
My social network workflow is a recursive loop of images and information that now passes between >> Feedly - Tweetdeck - Friendfeed << Feedly has become a keystone. - Brad Kligerman
For me, Greader is to Feedly as Twitter is to Friendfeed. Rss stream provides the data source, its inertia, but it requires another workflow modality to actually begin to understand its dense, complex meaning. - Brad Kligerman
completely automatic - if you're already signed in to those services. Freaked me out at first - but then loved it. - James Hull
use feedly since some monthes and LOVE it! Reading much more feeds since using it. Its much more "sexy" to go through the feeds now and easy share+discuss - natadd from twhirl
If you're looking for the Feedly mini-bar at the bottom, try opening a new window, I've found sometimes that helps to show it. Plus, I think it only comes up on feeds that you're subscribed to, or something like that. AGain, it's not perfectly clear. The buttons for sharing with friendfeed and twitter on individual posts should appear below each headline... - James Hull
How did I get 5 hours behind the eight ball on this?? Awesome, thanks Robert! - Phil Glockner
Feedly is excellent and the team work really hard to put in new features every week, and then release it! - Scott Davies
I can't quite see what is new, we already have add ons that do these things - what am I missing? - Iphigenie
will have to look it up, don't use firefox much anymore - but I had add ons for sharing, finding, looking up things like digg popularity - there's glue and Shareaholic and stumble upon and I forgot some of the other names. The only really new thing is the friendfeed stats, and I am of mixed feelings about it, because I wonder what effect tens of thousands of people doing a friendfeed search for every web page they visit will have on the cost and performance of the service? - Iphigenie
It's confused me so far - Richard A.
I'm reading this inside Feedly right now. Couldn't imagine keeping track of all the news without it. Hope it comes to Google Chrome. - Kris Haamer
@joelle. nice you meet you. Next time you are in the San Francisco Bay area and have some time, please give us a heads up and we will be happy to both show you how feedly is different and is built differently. - Edwin Khodabakchian
@edwin thanks for that - although I doubt I'll be in the area soon unless i get a surprise recruitment of some sort, being in Europe. Now I can see the interest of feedly as a well integrated and featured feed page/reader, I think, it's just some of the claims in this thread were a bit... much :D - Iphigenie
The "much" part is probably due to the euphoria of the launch (and Robert's touch!). But if you do a search for feedly on twitter you will notice that different people like/dislike different aspects of the experience (which allows us to continuously learn and improve it). If you have the chance, next time you have firefox up and running try the twitter sharing/re-tweeting feature of never know, you might be surprised! - Edwin Khodabakchian
Just because you have taken the time to explain I will install firefox and give it a good whirl :D - the only ff i have at the moment is ff2 on the linux mininote, so i need to install ff3 somewhere - Iphigenie
I've been trying it since yesterday. Impressed with flexibility and ease of use. Mini is great for forwarding to others. - Joe Buhler from twhirl
and of course the twitter posting. - Joe Buhler from twhirl
Twitter has fallen Angel Falls in my eyes, and not in the fun Base jumping way. - Richard A.
Didn't even realize it until the recommendations started popping up in Google Reader .... Cool Stuff. Congrats Edwin ... again - Charlie Anzman
“Flickr My Background” Puts Your Latest Flickr Photos as Your Twitter Background #web -
“Flickr My Background” Puts Your Latest Flickr Photos as Your Twitter Background #web
It keeps your Twitter bag fresh I guess. - Kol Tregaskes
Michael W. May
Little People - a tiny street art project: Tundra -
Little People - a tiny street art project: Tundra
Little People - a tiny street art project: Tundra
Show all
"Liverpool Street area, London" - Michael W. May from Bookmarklet
beyond tilt-shift! - Carlos Ayala
This guys stuff is something else. Love it. - Michael W. May
Chris Cotsapas
Q: any graph theory around weighting nodes AND edges (independently) in networks? Most seems to treat only edge weights.
can you be more specific ? most graph theory work (in biological networks) do not distinguish different node types. One lab that has been working in this direction is the lab of Luis Amaral. See for example this paper ( - Pedro Beltrao
Was thinking about weighting nodes (rather than node types in the sense of a bi/multi-partite graph) according to some biological data. Attaching GWAS association statistics to them, specifically. - Chris Cotsapas
Then no, I don't think I have come across something like that. - Pedro Beltrao
One possibility would be to do something like a two-step analysis of first defining modules and then doing the statistics on the nodes but I guess you already though of that :). Something along these lines - Pedro Beltrao
Thanks P - I'm actually trying to select the best path/subgraph based on both edge and node weights, so pre-specifying modules won't work. Ah well, another area of knowledge I will have to trailblaze :-) - Chris Cotsapas
@pedrobeltrao - just looked at the paper. I'm working with the first author on this project. Small world... - Chris Cotsapas
What is meant by best path? If you mean a path between two nodes then Djikstra's algorithm is what you want (shortest/cheapest path), though it only looks at edge weights (but a workaorund is to modify edge weights by their node wieghts). if you want to get shortest paths between all node pairs, use the Floyd Warshall algorithm. For subgraphs it depends on what type of subgraph you are looking for but clique detection algorithms might be useful. You can also look at work by David State regarding .... - Rajarshi Guha
approximate subgraph matching which may or may not be useful - Rajarshi Guha
@Rajarshi - I was also thinking about the possibility of changing edge weight according to the nodes weight but I am not sure that makes biological sense. Say for example that edge weights describe protein interaction propensity or degree of functional relatedness. Clustering this graph will give modules that are biologically relevant (ex.. complexes). - Pedro Beltrao
Lets say that node weight has no impact on the edge weight probability but we use it anyway and we have a new function describing edge weight that depends on the node weights. The modules that would come out of clustering the new graph might no longer be biologically relevant. - Pedro Beltrao
I guess if there is some sort of benchmark this function could be trained/tested. - Pedro Beltrao
Yes, combining edge and node weights may not be appropriate - depends on what the actual problem is. But to branch out a little, a chemical subgraph isomorphism algorithm does take into account nodes and edges - though here, it's not a matter of minimum cost. It either matches or not. In this sense the MOGA algorithm from David States allows fuzzy matches,so might be more suitable. Though I also wonder whether an algorthm that considers edge and node weight can be reduced to one that just considers... - Rajarshi Guha
...edge weights which combine the original edge weight and the node weights - Rajarshi Guha
@pedrobeltrao @rajarshi - thanks for the ideas. I'm looking to start from a given seed node and try to add nodes (or sets of nodes) by maximizing both edge weights (belief in the interaction) and node weights in some likelihood-esque function on the node weights. The two weight sets are completely independent. [does this make any sense?] - Chris Cotsapas
For small graphs, why not just brute force it?Or else for bigger graphs use a GA / SA to select nodes to add - basically it sounds like a feature selection problem. - Rajarshi Guha
I guess it does sounds like a feature selection prob. Brute force was my goto option - but I thought I'd see if more knowledgeable folks had more refined ideas :-) - Chris Cotsapas
not sure if this helps for your problem but perhaps this can be a source of inspiration: <> . They don't use weighted edges though. - Thomas Lemberger
@thomaslemberger - thanks, we've been poring over that pape for a little while... - Chris Cotsapas
Deepak Singh
The Train in Spain Replaces the Plane | Autopia from -
The Train in Spain Replaces the Plane | Autopia from
I looove train travel - Deepak Singh from Bookmarklet
I love to see a high speed track from Stockholm to Hamburg... - Egon Willighagen
Trains are soooo much better than planes! In fact, I'm on one right now, on my daily commute :) - Allyson Lister
Depends... the EuroStar (Bruxelles, London) has little leg space... less than RyanAir... but at least you have a chance for a table... in which case you can continue being present on FF :) - Egon Willighagen
I've gone skiing in the french alps (oooh!) on Eurostar a couple of times, and they've had great cheap first-class deals. Very nice! But even in standard I feel much less cramped than on a plane - maybe it's just an illusion... - Allyson Lister
Deepak Singh
Other ways to read this feed:Feed readerFacebook