Thanks Chris. This is great! Ideally, everything will be set up prior to the actual build meeting so all dependencies are in place. Let's work with Dan to make sure you have everything you need in the kbase image before the build, he's out on vacation this week, so next week. -----Original Message----- From: Christopher Henry [mailto:[email protected]] Sent: Sunday, July 29, 2012 1:46 AM To: Brettin, Thomas S. Cc: Scott Devoid; Paramvir Dehal; Gavin A Price; William Riehl; Paul Frybarger; Matt DeJongh; [email protected] Subject: KBase FBA modeling Hi Tom, So I'm starting to think about our manifest for the upcoming build. It's looking like all FBA modeling in KBase is going to go in a single repository. I set this up in git hub here: https://github.com/ModelSEED/KBaseFBAModeling Right now, the equivalent repo in KBase is called "model_api", which I believe we've all decided is a bad name. I think a name like "KBaseFBAModeling", "FBAModeling" would be more appropriate. We're prepping three repositories for deployment on build week: 1.) MFAToolkit: a c++ code that does all our flux balance analysis processing on the backend -https://github.com/ModelSEED/MFAToolkit 2.) ModelSEED: a massively trimmed down repo with the new Model SEED, which is the environment where many of the KBase FBA algorithms will be written (it can be considered at its heart, a perl object oriented API for FBA modeling) -https://github.com/ModelSEED/ModelSEED 3.) KBaseFBAModeling: the repo for the web services and IRIS front ends for all KBase FBA modeling services -https://github.com/ModelSEED/KBaseFBAModeling These three repos hold all the code that currently support the modeling services in IRIS. Our goal is by the time of the build, these repos will hold all services for reconciling phenotype data with models. The services in these repos will soon be applied to build draft models of all genomes in KBase CDM (genome_to_FBAmodel command). Draft models will be loaded in the CDM. This will be done before the build. Biochemistry is being loaded this upcoming week, along with phenotype data hopefully. I believe this biochemistry will include plant biochemistry, with support for plant modeling. So services we're targeting for the build: 1.) Plant model reconstruction (end of next week) 2.) Microbial model reconstruction (already in IRIS) 3.) Metagenome model reconstruction (this week) 4.) Model flux balance analysis including (already in IRIS) -flux variability analysis -single gene knockout -minimal media prediction -metabolic production analysis for non growing models -flux minimization -phenotype simulation (not sure the IRIS script supports this, but the web service does) 5.) Gapfilling of models (already in IRIS) 6.) Phenotype simulation and reconciliation (end of next week) Data we are targeting for the build: 1.) Biochemistry CDM (including plant biochemistry) 2.) Model CDM 3.) Phenotype CDM 4.) Document store for models and biochemistry Documentation is in progress... obviously, it's a lot to document, and these services are somewhat complex. Questions/comments welcome. Particularly, I want to know if you see something you were expecting us to produce but appears to be missing. Chris