The biostatistics groups at University of Helsinki focus on methods for computational modeling, probabilistic machine learning and inference. The groups are affiliated with the Department of Mathematics and Statistics, Faculty of Medicine, the Department of Biosciences and Institute for Molecular Medicine Finland and associated with the Centre of Excellence in Computational Inference COIN and the Helsinki Institute for Information Technology HIIT. Biostatistics is also among the strategic focus areas of University of Helsinki (See, e.g., the newly established Helsinki Institute of Life Sciences). More detailed information about our research, publications and researchers can be found under the principal investigators' pages.
The Bayesian statistics group (BSG), at the Department of Mathematics and Statistics, led by professor Jukka Corander focuses on statistical machine learning, inference for intractable models and on evolutionary epidemiology using genomics. We have recently spearheaded the use of Bayesian optimization to speed up inference for intractable simulator-based models by several orders of magnitude. BSG is part of the COIN Centre of Excellence in computational inference research (2012-2017) and also part of the COIN research programme (2016-) at Helsinki Institute of Information Technology (HIIT). The statistical methods introduced by the group have led to numerous important discoveries on the evolution, resistance, virulence and transmission of pathogenic bacteria and viruses.
Antti Honkela leads a research group working in Bayesian inference and machine learning at the Department of Mathematics and Statistics and the Department of Public Health.
We develop methods for efficient probabilistic inference in complex modelling problems. We develop models for genomic time series data using Gaussian processes and methods for quantitative analysis of sequencing data, including RNA-sequencing data and sequencing data from mixed bacterial samples. We also develop theory and methods for efficient differentially private Bayesian inference. We are a part of the COIN Centre of Excellence in computational inference research (2012-2017) and also a part of the COIN research programme (2016-) at Helsinki Institute for Information Technology (HIIT).
A new scientific modelling group, at the Department of Mathematics and Statistics, focuses on developing statistical methods for register-based studies, models for microbial metagenomics, real-world evidence and data for drug safety, and is led by professor Sangita Kulathinal. The main idea behind the use of health and social registers is to build individual-level transitions and trajectories and analyse them using statistical methods based on, for example, event history analysis and marked point processes, to address research questions related to early detection of diseases, health promotions and presonalised treatment paths. The group develops statistical methods needed in the microbial metagenomics research. Data from this field are typically large sparse matrix of count data. Event history analysis, marked point processes, copulas, generative models, regression and hierarchical models are among the methods that are being used/developed. The group collaborates closely with the National Institute for Health and Welfare (THL), Finland.
Very rapidly evolving technologies to read genomes have opened up exciting new possibilities to study relationships between the human genome and the observed traits, such as cholesterol levels, or susceptibility to disease, such as multiple sclerosis. We study, create and apply statistical methods needed in the modern human genetics research. The field provides challenging problems in large scale inference, high dimensional statistics as well as in updating traditional regression and classification models to the era of genomics. Efficient computation plays a crucial role in our work.
Samuli Ripatti is a Professor of Biometry at Faculty of Medicine and leads a research group working in Statistical and Translational Genetics at the Institute for Molecular Medicine Finland. They use genetic screens in Finnish population and disease samples to learn about disease mechanisms and risks. Cardiovascular diseases and risk factors are often used as models for genetics of complex diseases and statistical methods development. Our research is based on statistical genetics approaches in studying Finnish large-scale population and disease-based cohorts and family collections with genome-scale sequence data and rich sets of phenotypes.
Environmental sciences includes a broad range of scientific fields studying the environment and solutions to environmental challenges. Ecology studies the distribution and abundance of species, and their interactions with other species and the environment. Statistical inference and uncertainty estimation are essential for both fields to ensure that appropriate conclusions and decisions can be reached from experiments and observations. Our research interests span from the theoretical and computational statistics to applied ecology and environmental risk assessment and management.