Fast Integrative Factor Models: Applications from Nutritional Epidemiology to Cancer Genomics

The Bayesian Section of SSA is hosting the webinar: Fast Integrative Factor Models: Applications from Nutritional Epidemiology to Cancer Genomics presented by Dr. Alejandra Avalos-Pacheco.

Data-integration of multiple studies can be key to understand and gain knowledge in statistical research. However, such data present artifactual sources of variation, also known as covariate effects. Covariate effects can be complex, leading to systematic biases, that if not corrected could lead to unreliable inference. In this talk I will present novel sparse latent factor regression (FR) and multi-study factor regression (MSFR) models to integrate such heterogeneous data. The FR model provide a tool for data exploration via dimensionality reduction and sparse low-rank covariance estimation while correcting for a range of covariate effects. MSFR are extensions of FR that enable us to jointly obtain a covariance structure that models the group-specific covariances in addition to the common component. She will discuss the use of several sparse priors (local and non-local) to learn the dimension of the latent factors. Our approaches provide a flexible methodology for sparse factor regression which is not limited to data with covariate effects. Our models are fitted leveraging novel scalable EM and ECM algorithms as well as Variational Inference methods. She will present several examples, with a focus on bioinformatics applications. We show the usefulness of our methods in two main tasks: as an unsupervised dimension reduction task to give a visual representation of the latent factors of the data; and as a supervised tool to: (i) provide survival predictions leveraging the obtained factors, or (ii) obtain dietary patterns, associating each factor with a measure of overall diet quality related to cardiometabolic disease risk.

About the presenter:

Dr. Alejandra Avalos-Pacheco is an Universitätsassistentin (Assistant Professor non-tenure track) in the Research Unit of Applied Statistics at the Technische Universität Wien (TU Wien). Additionally, she holds an affiliation with Harvard University in the Harvard-MIT Center for Regulatory Science. Before joining TU Wien, Dr. Avalos-Pacheco served as a research fellow at the University of Florence. Prior to this, she held postdoctoral positions at Harvard University and the Dana-Farber Cancer Institute.


Dr. Avalos-Pacheco completed her PhD in Statistics through the joint CDT programme between the University of Warwick and the University of Oxford (OxWASP), focusing on statistical methods for genomic data analysis.  Her PhD thesis, supervised by Prof. David Rossell and Prof. Richard Savage, was granted the 2019 Savage Award in Applied Methodology. Such a prize is
conferred by the International Society of Bayesian Statistics (ISBA) and the American Statistical Association (ASA) Section on Bayesian Statistical Science (SBSS).

Dr. Avalos-Pacheco has actively contributed to the statistical community. She served as the chair of the junior section of the ISBA, known as j-ISBA, from 2021 to 2023.


Dr. Avalos-Pacheco is dedicated to enhancing understanding of real-world problems and facilitating fast, accurate decision-making. Her research interests lie in the application of statistical methodologies to medical challenges, particularly in the realm of cancer research. Leveraging Bayesian and probabilistic machine learning algorithms, she specialises in statistical methods for large heterogeneous data, emphasizing data integration. Her research interests span high-dimensional inference, applied Bayesian statistical modeling, dimensionality reduction, heterogeneous data integration, graphical models, and clinical trials.