>_rg::lab

Open positions

Raphael Gottardo — Thu, 26 Aug 2010 15:13:27 +0000

The rglab is currently looking for a scientific programmer and a postdoctoral fellow. We seek people with strong computational skills and interests in immunology and/or vaccine development. Please visit the Jobs page for more info. Note that the Fred Hutchinson Cancer Research Center is constantly ranked as one of the best places to work in academia.

The rglab at the FHCRC

Raphael Gottardo — Tue, 24 Aug 2010 20:44:55 +0000

The rglab has just moved to the Fred Hutchinson Cancer Research Center (FHCRC). We are now part of the Vaccine and Infectious Disease Division at the FHCRC. We are located in the Arnold building shown on the left. We will be working on computational problems related to immunology and vaccine development. If you need to contact us, please use the contact page.

Welcome to the rglab!

Raphael Gottardo — Tue, 24 Aug 2010 17:54:22 +0000

We conduct research in computational biology and statistical genomics with applications to high throughput biological assays and immunology. The rglab is located within the Vaccine and Infection Disease Division at the Fred Hutchinson Cancer Research Center in Seattle.

Flow cytometry special issue

Raphael Gottardo — Tue, 24 Aug 2010 17:00:10 +0000

Our special issue on “Recent Bioinformatics Advances in the Analysis of High Throughput Flow Cytometry Data” has been published in Advances in Bioinformatics. This includes our own work on flowMerge, shown on the left.

Let the Data Flow

Raphael Gottardo — Sat, 10 Jul 2010 00:37:06 +0000

Last month’s issue of The Scientist had an interesting article about flow cytometry data analysis, and in particular about software and tools available for data analysis. The good news is that the article mention Bioconductor. The bad news is that it was almost a simple footnote, to quote the article it said:

“Don’t forget freeware. WinMDI is another popular choice, although it is no longer supported. “It does very basic analysis, and that’s maybe all you want initially,” Davies says. WinMDI is not compatible with some of the newer flow cytometry instruments. Other freebies include FlowingSoftware, WEASEL and open-source flow cytometry tools within Bioconductor.”

So I would say don’t forget the freeware but more importantly think open source! Open source is the ONLY way to go for reproducible research! How can you expect to reproduce someone’s result if you don’t have the software or can’t afford it?

Bioconductor (and R) is open source, free, and cross platform, so you won’t have any problems reproducing the results of your colleague even if you don’t have a Mac or a Windows machine, or whatever. Of course, the fact that the article mentions Bioconductor is already a great start as we started our FCM project a couple of years ago only. Let’s hope for more next time!

Bioconductor offers many great tools for flow data analysis and as far as I am concerned is far ahead of competitors in the field of automated data analysis. If you want to learn more about flow cytometry data analysis you could think about attending the Bioconductor flow cytometry sessions: https://secure.bioconductor.org/BioC2010/labs.php

Here is the link if you’re interested: http://www.the-scientist.com/article/display/57473/

Unfortunately it’s paid subscription only.

rMAT and openMP

Raphael Gottardo — Sat, 03 Jul 2010 01:12:14 +0000

So we have been working on enabling openMP for rMAT. So far we have only been using Grand Central Dispatch for parallel computation as we mainly use Mac OS X 10.6. This has been available in rMAT and rGADEM for quite a bit already, and it works great. This being said, we know that not everyone uses Apple machines, so we wanted to have a similar version for *nix and Windows users. Our implementation will be based on openMP, which is available (or can be made available) on most machines. Today I am pleased to announce that rMAT version 2.6 or greater supports this already. We are now working on rGADEM, so please stay tuned. We have included a configure script so that the package will try to detect openMP before you install it and set up the corresponding flags. To make a long story short, rMAT with openMP is available from BioC 2.7 and soon rGADEM will be too. This work has been done thanks to Davor Cubranic at UBC who will be doing a bit of work for us in the next few months.

chip-chip and chip-seq course at CSHL

Raphael Gottardo — Sat, 03 Jul 2010 00:28:00 +0000

I recently gave two lectures on chip-chip and chip-seq at the Cold Spring Harbor Labs as part of the “Integrative Statistical Analysis of Genome Scale Data” course. During my lectures and labs I have covered various aspects of the analysis of chip data going from raw data to enriched regions and de novo motifs. In particular, I talked about our R packages: rMAT, PICS, rGADEM and MotIV, all available from Bioconductor. Overall, the course was very good and the students were great!

For the purpose of this course I have put an R package containing all necessary data and a complete detailed vignette that was used for the labs. I thought that this could be a good ressource for other people, so I have decided to write this post with links to the R package. You can easier download the package as a tar ball here, or from the following svn repository: http://svn.rglab.org/packages/chipData/ (it’s public).

Even though downloading from the svn requires more bandwidth as the data are not zipped, I recommend it over the archive as I might update the files in the future. Once you check it out once, you will always be able to update your directory easily without downloading everything again.

Feel free to share this with anyone who might be interested.

Enjoy! As always, comments and suggestions are welcome!

Congratulations to Marie-Pier

Raphael Gottardo — Tue, 22 Jun 2010 01:23:42 +0000

Congratulations to Marie-Pier who won an award for best poster presentation during the IRCM research day. Please visit this page for more details.

Well done Marie-Pier!

Paper selected as JCGS highlight of 2010

Raphael Gottardo — Tue, 22 Jun 2010 00:31:16 +0000

The paper “Combining Mixture Components for Clustering“, by Jean-Patrick Baudry, Adrian Raftery, Gilles Celeux, Raphael Gottardo and Kenneth Lo, was selected as a Highlight of 2010 in the Journal of Computational and Graphical Statistics.

We have used this methodology for clustering flow cytometry data, as implemented in flowMerge.

GSL and 64 bit windows

Raphael Gottardo — Mon, 31 May 2010 15:44:37 +0000

Good news for 64-bit Windows users! Thanks to Arnaud Droit (and to Brian Ripley and Uwe Ligges who provided essential information to help solving this issue), a 64-bit version of the GSL (GNU Scientific Library) is now available on our wiki. You will find binaries for 32 and 64 bit windows as well as instruction to build it yourself.

Thanks to these Binaries and the work of Herve Pages from the BioC team, all of our R packages (PICS, rMAT and MotIV) requiring GSL are now available on both 32 and 64 bit binaries for windows. These packages can be downloaded from Bioconductor. So please try them out!