Biostatistics & Informatics


Develops novel statistical & computational methodology to address the challenging data analysis problems arising in modern biomedical, genomics, health & environmental studies.

Theme Leader: Professor Marta Blangiardo (ICL) 

Key Projects

  • Integration of data sources (e.g. population based registries & cohorts/surveys): provide accurate confounder adjustment of the association between exposures/risk factors & health outcomes at the small-area level, dealing with the issue of missing data. (MRC Methodology funded grant). Extend the developed framework to deal with categorical/continuous exposure & with multi-pollutants.
  • Misalignment in exposure and health data: develop a Bayesian spatio-temporal model to combine exposure data available at point locations and grid & health outcome available at administrative areas. Applications include e-health data & routinely available datasets (hospital episode statistics & mortality registries), to investigate the association with noise & air pollution.
  • Bioinformatics applications: multivariate methods for biomarker discovery & variable selection, nonparametric Bayesian approaches for clustering & data integration of heterogeneous omics data & classical clinical covariates, gene & mediation network modelling.
  • Space/time disease mapping: to develop robust & flexible models for disease surveillance able to distinguish a common temporal trend from area-specific (uncommon) ones; to evaluate effects of policies (e.g. European Directive for Incinerators) taking into account spatial & temporal dependencies.