My code is hosted on GitHub
You'll find there:
- A fixed-sized parametric maxflow implementation
- Ongoing work on sample-specific co-expression networks
- Ongoing work on network-guided feature selection
Multitask Lasso with task descriptors
A multi-task Lasso approach that makes use of task descriptors. The code, developed by Víctor Bellón, is available on GitHub.
Víctor Bellón, Véronique Stoven, and Chloé-Agathe Azencott. Multitask feature selection with task descriptors, Pacific Symposium on Biocomputing, 2016, 21:261—272 [pdf].
SConES: electing nected xplanatory NPs
SConES is a network-guided approach for analyzing genome-wide data. It allows for the discovery of multiple genetic loci that are maximally associated with a phenotype, while tending to be connected on a given biological network. This network can be constructed for example from a gene-gene interaction network (based on proximity), or in any way such that you believe that neighboring SNPs should tend to be selected together.
Chloé-Agathe Azencott, Dominik Grimm, Mahito Sugiyama, Yoshinobu Kawahara, and Karsten M. Borgwardt. Efficient network-guided multi-locus association mapping with graph cuts, Bioinformatics, 2013, 29(13): i171—i179 DOI: 10.1093/bioinformatics/btt238 (Open Access)
Matlab code developed by Dominik Grimm, Yoshinobu Kawahara and myself is available on GitHub.
SConES is also available as part of EasyGWAS, a framework for the analysis and meta-analysis of GWAS data (with Python interfaces).
R code for Multi-Scones, developed by Mahito Sugiyama, is available on GitHub.
GLIDE: PU-based near etection of pistasis
GLIDE is a GPU-based approach for the detection of epistasis in genome-wide data. It allows for the systematic computation of a linear regression between pairs of genetic loci and a phenotype.
Tony Kam-Thong*, Chloé-Agathe Azencott*, Lawrence Cayton, Benno Pütz, André Altmann, Nazanin Karbalai, Philipp G. Sämann, Bernhard Schölkopf, Betram Müller-Myhsok, and Karsten M. Borgwardt. GLIDE: GPU-based linear regression for the detection of epistasis. Human Heredity, 2012, 73:220—236. DOI: 10.1159/000341885 [pdf]
CUDA code (for execution on NVIDIA GPUs) developed by Tony Kam-Thong and myself is available on GitHub.
An additional How To as well as (Python and bash) scripts for working with GLIDE, that I developed in the context of a case study, are available on GitHub.