MLPM Project


From Genetic Data to Medicine

Over the last decade, enormous progress has been made on recording the health state of an individual patient down to the molecular level of gene activity and genomic information – even sequencing a patient’s genome for less than 1000 dollars is no longer an unrealistic goal. However, the ultimate hope to use all this information for personalized medicine, that is to tailor medical treatment to the needs of an individual, remains largely unfulfilled. To turn the vision of personalized medicine into reality, many methodological problems remain to be solved: there is a lack of methods that allow us to gain a causal understanding of the underlying disease mechanisms, including gene-gene and gene-environment interactions. Similarly, there is an urgent need for integration of the heterogeneous patient data currently available, for improved and robust biomarker discovery for disease diagnosis, prognosis and therapy outcome prediction.

Bringing together Machine Learning and Statistical Genetics

The field of machine learning, which tries to detect patterns, rules and statistical dependencies in large datasets, has also witnessed dramatic progress over the last decade and has had a profound impact on the Internet. Amongst others, advanced methods for high-dimensional feature selection, causality inference, and data integration have been developed or are topics of current research. These techniques address many of the key methodological challenges that personalized medicine faces today and keep it from rising to the next level. Despite this rich potential of machine learning in personalized medicine, its impact on data-driven medicine remains low, due to a lack of experts with knowledge in both machine learning and in statistical genetics. Our ITN aims to close this gap by bringing together leading European research institutes in Machine Learning and Statistical Genetics, both from the private and public sector, to train 14 early stage researchers. The project is sponsored by EU FP7-PEOPLE Project Ref 316861 and is a collaboration with Magnus Rattray of University of Manchester, Karsten Borgwardt of MPI for Intelligent Systems, Bernhard Schoelkopf of MPI for Intelligent Systems, Bertram Mueller-Myhsok of MPI for Psychiatry, Volker Tresp of Siemens AG, Felix Agakov of Pharmatics LTD, Kristel Van Steen of Universit'e de Li`ege, Jean Philippe Vert of Mines ParisTech (Armines), Florence Demenais of INSERM, Fernando Perez Cruz of Universidad Carlos III de Madrid, Joaquin Dopazo of Principe Felipe Centro de Investigacion and Gunnar Raetsch of Memorial Sloan-Kettering Cancer Center.

Personnel from ML@SITraN


The following software has been made available either wholly or partly as a result of work on this project:- GPy GPy: Gaussian process modelling framework in Python


The following conference publications were made associated with this project.

J. Hensman, M. Zwiessele and N. D. Lawrence. (2014) “Tilted variational Bayes” in S. Kaski and J. Corander (eds) Proceedings of the Seventeenth International Workshop on Artificial Intelligence and Statistics, JMLR W&CP 33, Iceland, pp . [Software][Google Scholar Search]


We present a novel method for approximate inference. Using some of the constructs from expectation propagation (EP), we derive a lower bound of the marginal likelihood in a similar fashion to variational Bayes (VB). The method combines some of the benefits of VB and EP: it can be used with light-tailed likelihoods (where traditional VB fails), and it provides a lower bound on the marginal likelihood. We apply the method to Gaussian process classification, a situation where the Kullback-Leibler divergence minimized in traditional VB can be infinite, and to robust Gaussian process regression, where the inference process is dramatically simplified in comparison to EP.\ \ Code to reproduce all the experiments can be found at

R. Andrade-Pacheco, J. Hensman and N. D. Lawrence. (2014) “Hybrid discriminative-generative approaches with Gaussian processes” in S. Kaski and J. Corander (eds) Proceedings of the Seventeenth International Workshop on Artificial Intelligence and Statistics, JMLR W&CP 33, Iceland, pp . [Software][Google Scholar Search]


Machine learning practitioners are often faced with a choice between a discriminative and a generative approach to modelling. Here, we present a model based on a hybrid approach that breaks down some of the barriers between the discriminative and generative points of view, allowing continuous dimensionality reduction of hybrid discrete-continous data, discriminative classification with missing inputs and manifold learning informed by class labels.

The following publications have provided background to our work in this project.

J. Hensman, N. D. Lawrence and M. Rattray. (2013) “Hierarchical Bayesian modelling of gene expression time series across irregularly sampled replicates and clusters” in BMC Bioinformatics 14 (252) [DOI][Google Scholar Search]


\textbf{Background}\ \ Time course data from microarrays and high-throughput sequencing experiments require simple, computationally efficient and powerful statistical models to extract meaningful biological signal, and for tasks such as data fusion and clustering. Existing methodologies fail to capture either the temporal or replicated nature of the experiments, and often impose constraints on the data collection process, such as regularly spaced samples, or similar sampling schema across replications.\ \ \textbf{Results}\ \ We propose hierarchical Gaussian processes as a general model of gene expression time-series, with application to a variety of problems. In particular, we illustrate the method’s capacity for missing data imputation, data fusion and clustering.The method can impute data which is missing both systematically and at random: in a hold-out test on real data, performance is significantly better than commonly used imputation methods. The method’s ability to model inter- and intra-cluster variance leads to more biologically meaningful clusters. The approach removes the necessity for evenly spaced samples, an advantage illustrated on a developmental Drosophila dataset with irregular replications.\ \ \textbf{Conclusion}\ \ The hierarchical Gaussian process model provides an excellent statistical basis for several gene-expression time-series tasks. It has only a few additional parameters over a regular GP, has negligible additional complexity, is easily implemented and can be integrated into several existing algorithms. Our experiments were implemented in python, and are available from the authors’ website:

N. Fusi, C. Lippert, K. Borgwardt, N. D. Lawrence and O. Stegle. (2013) “Detecting regulatory gene-environment interactions with unmeasured environmental factors” in Bioinformatics 29 (11), pp 1382–1389 [DOI][Google Scholar Search]


\textbf{Motivation}: Genomic studies have revealed a substantial heritable component of the transcriptional state of the cell. To fully understand the genetic regulation of gene expression variability, it is important to study the effect of genotype in the context of external factors such as alternative environmental conditions. In model systems, explicit environmental perturbations have been considered for this purpose, allowing to directly test for environment-specific genetic effects. However, such experiments are limited to species that can be profiled in controlled environments, hampering their use in important systems such as human. Moreover, even in seemingly tightly regulated experimental conditions, subtle environmental perturbations cannot be ruled out, and hence unknown environmental influences are frequent. Here, we propose a model-based approach to simultaneously infer unmeasured environmental factors from gene expression profiles and use them in genetic analyses, identifying environment-specific associations between polymorphic loci and individual gene expression traits.\ \ \textbf{Results}: In extensive simulation studies, we show that our method is able to accurately reconstruct environmental factors and their interactions with genotype in a variety of settings. We further illustrate the use of our model in a real-world dataset in which one environmental factor has been explicitly experimentally controlled. Our method is able to accurately reconstruct the true underlying environmental factor even if it’s not given as an input, allowing to detect genuine genotype-environment interactions. In addition to the known environmental factor, we find unmeasured factors involved in novel genotype-environment interactions. Our results suggest that interactions with both known and unknown environmental factors significantly contribute to gene expression variability.\ \ \textbf{Availability}: Software available at\ \ \textbf{Contact}:,

J. Hensman, M. Rattray and N. D. Lawrence. (2012) “Fast variational inference in the conjugate exponential family” in P. L. Bartlett, F. C. N. Pereira, C. J. C. Burges, L éo. Bottou and K. Q. Weinberger (eds) NIPS, . [PDF][Google Scholar Search]

J. Hensman, N. Fusi and N. D. Lawrence. (2013) “Gaussian processes for big data” in A. Nicholson and P. Smyth (eds) Uncertainty in Artificial Intelligence, AUAI Press, . [PDF][Google Scholar Search]

N. Fusi, O. Stegle and N. D. Lawrence. (2012) “Joint modelling of confounding factors and prominent genetic regulators provides increased accuracy in genetical genomics studies” in PLoS Computat Biol 8, pp e1002330 [Software][PDF][DOI][Google Scholar Search]


Expression quantitative trait loci (eQTL) studies are an integral tool to investigate the genetic component of gene expression variation. A major challenge in the analysis of such studies are hidden confounding factors, such as unobserved covariates or unknown subtle environmental perturbations. These factors can induce a pronounced artifactual correlation structure in the expression profiles, which may create spurious false associations or mask real genetic association signals. Here, we report PANAMA (Probabilistic ANAlysis of genoMic dAta), a novel probabilistic model to account for confounding factors within an eQTL analysis. In contrast to previous methods, PANAMA learns hidden factors jointly with the effect of prominent genetic regulators. As a result, this new model can more accurately distinguish true genetic association signals from confounding variation. We applied our model and compared it to existing methods on different datasets and biological systems. PANAMA consistently performs better than alternative methods, and finds in particular substantially more trans regulators. Importantly, our approach not only identifies a greater number of associations, but also yields hits that are biologically more plausible and can be better reproduced between independent studies. A software implementation of PANAMA is freely available online at