Composite likelihood is a likelihood modification useful in instances where maximum likelihood estimation (MLE) is computationally infeasible. The first part of the talk focuses on discrete latent variable models for two-way data arrays, in which MLE is intractable due to the complex structures of the models. We construct composite likelihood as a computationally tractable alternative to the full likelihood of our models, and discuss the performance of our methods via simulations and applications to genomic data.
With over 20 million formalin-fixed, paraffin-embedded (FFPE) tissue samples archived each year in the US alone, archival tissues remain a vast and under-utilized resource in the genomic study of cancer. The focus of the second part of this talk will be on open methodological challenges relate to RNA expression profiling in FFPE tissue samples.