After long time analyzing bioinformatics scientist behaviour, I finally figured out that no real effort is done to let computational biologist interact with each other, a simple observation can serve as a proof of my statement, there is no collaborative platform for sharing source code dedicated to bioinformatics, a simple but very important issue that could help scientists develop even 10 times faster, by providing "patterns" of what other scientist have already done.
Developers and communities of Bio* projects behave like a closed group in regards to others, a considerable effort and useful tool such as mailing lists certainly provides communities with useful help, but there is no real effort to publish solved source code after they are examined and corrected by experienced programmers, resulting to make developers and mailing list admin to answer sometimes the same problem twice and even much more. (I see sometimes list admins refering to a previous post or thread, that's a good proof)
Starting from that point, I tried to see if such initiative was already taken, none of bioconductor, bioperl, biojava, biopython communities have a resources website storing high quality source codes and case study codes (of course I am not talking about all demo codes, cook books, or illustrations that constitute the developers starter kits to let scientist adapt a Bio* project to their need)
As I used to work with bioperl, biojava especially and now working with Bioconductor and R since I switched my research to the epigenomics side, I asked some people in the BioC list if they are willing to see a collaborative platform for people wanting to share their source codes, and the result was very appealing :
How can you describe your R/Bioconductor learning curve ?
Slow 12%, Intermediate 53%, Fast 35%
Do you easily find R/Bioconductor codes, scripts, workflows ?
Yes 35%, No 65%
How do you describe your experience in R/Bioconductor ?
Beginner 18%, Intermediate 53%, Experienced 29%
Do you think an R/Bioconductor community could help to enhance your develoment ?
Yes 82%, No 0%, I don't know 18%
These stats illustrate the importance of having such collaborative platform website allowing bioinformaticians and computational biologists REALLY and EFFECTIVELY collaborate and enhance their developments.
I already developed a prototype (90% of the development is already done), a possible annoucement is possibly planned for the last week of January or early February, although the website is already hosted and effectively running, I will need help to test it and report eventual bugs and problems. The website is intuitive, easy to use and well designed, it recquire a free subscription, users will edit their profile and they are offered a "Submitter tool" to post their source code in three simple steps :
I will need contributors and beta testers for this website and I will happy to count you among the beta testers of this new developed tool.
Please let me know as soon as possible if you are willing to participate in the website evaluation and testing sending as much source codes as you can.
Kind regards and Happy new year.
Radhouane
PS : Some biojava and bioperl users proposed to extend the website to java and perl, what I in fact did adding also biopython.
If you are interested in sharing your source codes please contact me asap through this website and put as subject : Sharing Source Codes
•
written
by
Radhouane Aniba
I disagree that bioinformaticians operate as closed groups, since you can find the source-code of many projects on either SourceForge, GitHub, or privately hosted. This works fairly well, since at least SourceForge and GitHub give you the possibility to track the development of the software as well as to store documentation in form of web-pages or wikis.
Joachim I never said it will be a closed group, that's completely the opposite of what I described in the text, sorry if you missunderstood the message.
In my opinion it will essential to be able to interact with the said source codes in a code versioning system (git, mercurial etc). For example the way gist works:
Code versioning system is what I am working on actually as a feature of the website, Thanks istvan
I think that your opening statement that "no real effort is done" comes across as rather dismissive of prior work such as bioinformatics.org.
I've upvoted Joachim's answer because I also disagree with the assertion that bioinformaticians operate as closed groups. In particular, I have never found this of the developers and communities of the Bio* projects. In fact, quite the reverse; they have always most welcoming to me as a developer over the past ten years. I can say the same for Ensembl.
To answer the question, yes I do have source to share. You can find it on github, linked from my profile. Centralisation of repositories is unnecessary; some VCS are distributed anyway and for the rest we have links. This is the web, right?
Hi Keith, Again I did not wanted to minimize the work of other people like bioinformatics.org or any other people trying to make things better for the community, I am the first to clap my hands for their efforts. I just presented an initiative for something that I personnaly think new (Not sharing Projects, but sharing snippets). Why making a website for bioinformatics developers is considered as making a closed group ? Is that the case for bioinformatics.org? defenitly not.
Log in to answer this question.
Or you can contact me via this URL
http://tinyurl.com/4bm6x5g
Have a nice day,
Radhouane
I don't see a concrete question here ...
Who own/support/fund this proposed network ?
I (Radhouane Aniba, Bioinformatics Research Associate at University of maryland) am supporting this network personnaly and would be grateful if people could help to improve the developing of such initiative by providing ideas and actively contributing and interacting
What will be the single advantage of sharing source code in the proposed network than existing resources like SourceForge, GitHub or private hosting as mentioned but Joachim & Istavan ?
What will be the single advantage of sharing source code in the proposed network than existing resources like SourceForge, GitHub or private hosting as mentioned by Joachim & Istavan ?
Code centralisation by Bioinformatics speciality. In other terms, Sourceforge, Github or private hosting is proposing "Projects Hosting and Versionning" which is great and very useful. What I would like to offer is "Project case studies Hosting" using Bio* packages. Let's take Biojava, What I may do with it is probably very different from what you can do, but both of our codes is certainly useful for others. These projects comes with extended and rich documentation and codes through their wikis and websites but what people do with that is as far as I know never stored before.