Sunday, 19 June 2011

Gingko-CADx - or - Debian Med is not only about bioinformatics

There is a new package in town, namely Gingko-CADx. This is all about viewing and (for those who want to) also about programming with image data from medical devices. And there are only a handful of packages that render it so obvious why Open Source is about freedom. A couple of years ago I was not so very concerned when I could not immediately read the image data from a befriended clinical institute. My personal eye opener was when I learned that the local Gynecologist could not read the 3D image data but relies on his ultra sound experience and the summary of the Radiologist. I am not so worried about him deciding not to look at it, it is more that this doctor cannot freely make this decision. He does not have the software. Well, he does, now, or to be more exact .. I should go and bring it.

Also in my field, the investigation of complex diseases, image data and their automated analysis is gaining ground. There are movies on the activity of animals, videos counting migrating cells, substances traced in the body ... the more molecular the better. This way, one aims at characterising intermediate phenotypes that are more strongly coupled to genotypic variations, and thus one hopes to better explain the complete disease phenotypes with those additional insights. This is not an academic exercise, but the very start of an understanding that different patients need different treatments. And we will understand better, in what ways the animal models of diseases differ from what one can observe in humans.

So, Open Source once again brings us better communication between doctors, patients and doctors, and between either and statistical geneticists. Yippee.

Friday, 3 June 2011

Community efforts towards research on EHEC

Northern Germany these days experiences many hundreds of infections with severest clinical symptoms by a new strain of E. coli bacteria EHEC subtype. The genome has recently been sequenced only to find out that it is something novel, featuring the EHEC surface markers with something from African strains. No, antibiotics don't work, yes, the number of cases are still increasing, no, the source has not been found.

What would be great to have readily available for everyone are now tools to present some comparative genomics of that strain against the other strains of E. coli sequenced so far. I was very happy to read here about collaboratively analysing the raw sequencing data. The next step would then be to look at strain differences and interpret them - on the molecular level and in the clinical context. So, whatever tools we could help promoting to the more clinical researchers, they might be used for a while.

I am not completely sure about what Debian Med could help best with. So, much in line of some earlier blog entry on "doing things with Debian Med" I ask just anyone in contact with researchers working with E. coli (or other pathogenic bacteria) to report on what tools they would like to have available through us or what tools they are already using that Debian Med does not yet provide (so other can have it more easily).

Update [10th of June]: Through the Bioinformatics Computation group at LinkedIn I was pointed to It looks very nice though I have not any idea yet about how to work with it. And I learned about yet another tool, RAST, standing for 'Rapid Annotation using Subsystem Technology'. More comments, please.

Update [14th of June]: A group in Saarbuecken focuses on regulatory transcriptional motifs of EColi.