"Official" September 2014 meetup

Agenda

6:30 - Pizza and networking
7:00 - Announcements
7:10 - Tyler Backman: Systems Biology Drug Discovery with R
7:30 - Lightning talks
7:30 - Antonio Piccolboni: 10 eigen maps of the United States of America
7:45 - Dennis Noren: R Shiny and Baby Names
8:00 - Dave Deriso: R on Heroku: Tales from the Production Battlefield
8:15 - Nicole White: R and Neo4J for managing highly-connected data sets:
8:30 - Vince Scopino: TBD

------------------------------------------------------------------

Systems Biology Drug Discovery with R

Abstract: Identification of small molecules likely to have a very specific biological effect is a major challenge in drug discovery. Bioassay experiments, which assess the behavior of small molecules in a specific biological context are a key method for identifying potential drug candidates. To date hundreds of thousands of bioassay experiments have been published in public databases yet comparing data across numerous heterogenous experiments is currently a difficult task for the computational biologist; thus far analysis has been limited to small subsets of these data. I have developed a suite of R language software tools which systematically analyze these data to identify target selective drug candidates, and assess the drugability of potential protein targets.

Bio: Tyler Backman develops software to identify molecules likely to treat human disease. He is currently a Biomedical Engineering PhD candidate at UC Riverside.

Antonio Piccolboni - 10 eigen maps of the United States of America

The inspiration for this talk came from a NYT article (http://www.nytimes.com/2014/06/26/upshot/where-are-the-hardest-places-to-live-in-the-us.html) that shows a map of "hard life" in the United States. But what happens if instead of picking 6 variables we picked 6000? What if instead of making up a statistics out of thin air we applied standard methods? The result are ten eigenmaps of the United States.

------------------------------------------------------------------

Dennis Noren - R, Shiny and Baby Names

I use the R Shiny package for an application to explore the babynames package as captured by Hadley Wickham from the US Social Security Administration database. My starting point is a shiny tutorial presented by Garrett Grolemund, and I have added functionality for selection, comparison, time series smoothing, and sorted lists. This demonstrates concepts in reactivity, plot and table rendering, data.table, and compact UI design.

----------------------------------------------------------------------

Dave Deriso - R on Heroku: Tales from the Production Battlefield
In this talk I will discuss a few hacks that will get you up and running with R and Ruby on Rails using a standard Heroku dyno. Dave Deriso is a graduate student in Stanford's Institute for Computational and Mathematical Engineering.

---------------------------------------------------------------------

Nicole White - R and Neo4J for managing highly-connected data sets

R and igraph alone comprise an incomplete toolset for network analysis. Efficiently managing complex, highly-connected datasets requires data persistence. Complete your toolset with Neo4j, a graph database, and RNeo4j, Neo4j's R driver. Persist your data in Neo4j, use RNeo4j to coerce mathematical graphs out of your data, and use igraph for executing graph algorithms. Watch a short demonstration of this store, query, and analyze approach with the added benefit of RNeo4j.