Toby Dylan Hocking

Toby Dylan Hocking
INRIA - Sierra team for machine learning research
Laboratoire d'Informatique de l'Ecole Normale Supérieure
23, avenue d'Italie
CS 81321
75214 Paris Cedex 13
+33 1 39 63 54 99
toby's email

I am a doctoral student of statistical machine learning applied to bioinformatics. My advisors are Francis Bach and Jean-Philippe Vert. My CV/résumé is available in English and French, and my list of publications is available in BibTeX. My principal research interests are

Statistical learning methodology, especially encoding prior knowledge using structured sparsity-inducing norms, and unsupervised methods such as clustering.
Applications of statistical learning in bioinformatics, especially using interpretable models to better understand the genetic factors involved in cancer biology.
Statistical software development, for which I have written some free software for clustering, bioinformatics, population genetics, visualization, and documentation generation for the R project.
Visualization of high-dimensional data using techniques such as animation and interactivity.

Publications, talks, posters [RSS feed]

	Recent advances in direct labeled graphics [source]. Invited talk for semin-r, the Paris R user group (2012).
	Hocking TD, Schleiermacher G, Janoueix-Lerosey I, Delattre O, Bach F, Vert J-P. Learning smoothing models of copy number profiles using breakpoint annotations. [14-page HAL preprint] [Python annotation GUI] [Model training R code] [Annotated neuroblastoma copy number data] [Evry invited talk slides source]
	Adding direct labels to plots, which was awarded Best Student Poster [source], and Fast, named capture regular expressions in R 2.14, a 5-minute lightning talk [source], for useR 2011.
	Collaborative R package development using R-Forge and updated Sustainable R package development using documentation generation [source]. Invited talks for a seminar on bioinformatics at the Institut de Biologie de Lille, 9 June 2011.
	Hocking TD, Joulin A, Bach F and Vert J-P. Clusterpath: An Algorithm for Clustering using Convex Fusion Penalties. In proceedings of the 28th International Conference on Machine Learning (2011). [8 page conference paper] [HAL preprint] [ICML poster] [Digiteo poster] [slides] [video] [L[12] solvers in R] [L∞ solver in MATLAB] [source]
	Sustainable package development using documentation generation. Invited talk for semin-r, the Paris R user group (2010).
	Gautier M, Hocking TD, Foulley J-L. A Bayesian Outlier Criterion to Detect SNPs under Selection in Large Data Sets, PLoS ONE 5(8). Work related to my masters thesis (2010).
	Visualizing multivariate data using lattice and direct labels. Invited talk for semin-r, the Paris R user group (2009).
	Identifying genetic loci under selective pressure using a hierarchical Bayesian model with a posterior predictive p-value classifier. Masters thesis for Pierre and Marie Curie University, with advisors Mathieu Gautier and Jean-Louis Foulley of INRA Jouy-en-Josas (2009).
	Sublogo dendrograms: visualizing correlation in biological sequence motifs. Poster for useR! 2009.
	Perez et al. Establishment of HIV-1 resistance in CD4+ T cells by genome editing using zinc-finger nucleases. Nature Biotechnology 26, 808 - 816 (2008). I did the analysis of the evolution of the HIV V3 loop after the cells were treated with ZFNs, resulting in Supplementary Figures 4c and 4d.
	Doyon Y, McCammon JM, Miller JC, Faraji F, Ngo C, Katibah GE, Amora R, Hocking TD, Zhang L, Rebar EJ, Gregory PD, Urnov FD, Amacher SL. Heritable targeted gene disruption in zebrafish using designed zinc-finger nucleases, Nature Biotechnology, 26, 702-708 (2008). I did the experiments that characterized the DNA-binding specificity of the ZFNs used for the knock-outs, described in Supplementary Figure 11.
	Chromosomal copy number analysis using SNP microarrays: an extension of RLMM with a binomial test statistic. Undergraduate thesis at UC Berkeley with advisor Terry Speed (2006).