Dr Miguel Lacerda

Adjunct Senior Lecturer

Research Interests: 

Bioinformatics; Stochastic processes

Courses Taught

  • STA1006S – Statistics for Mathematical Disciplines
  • STA2004F – Statistical Theory and Inference
  • STA3045F – Advanced Stochastic Processes
  • Theory of Statistics (Honours)

Biographical Information

I am a former graduate student from the University of Cape Town, having obtained my bachelors degree in Business Science in December 2005 and my masters degree in Statistical Sciences in June 2008. Both degrees were awarded with distinction. I lectured on a contract basis in the Department of Statistical Sciences for two years while completing my masters, before moving to the National University of Ireland, Galway to pursue my doctorate in Bioinformatics under the supervision of Prof Cathal Seoighe. My PhD thesis investigated how HIV evolves in order to avoid the host’s immune response and introduces novel statistical methods for identifying regions of the viral genome that are relevant for vaccine design. I assumed a permanent position of lecturer in the Department of Statistical Sciences at UCT in September 2009.

Research Interests

I am primarily interested in statistical problems in bioinformatics, with a focus on computational molecular evolution, phylogenetics and population genetics. I am particularly interested in how host-pathogen interactions shape the evolution of the Human Immunodeficiency Virus (HIV), allowing the virus to escape the immune response and drug therapy.

My recent work involved the development of phylogenetic models that allow the stochastic process of molecular evolution to depend on environmental variables, such as genetic markers of the host's immune system in the case of viral evolution. Under the supervision of Prof Cathal Seoighe at the National University of Ireland, Galway, I have developed a phylogenetic hidden Markov model (phylo-HMM) to identify regions of the HIV genome (called epitopes) that illicit an adaptive immune response and which may be relevant for vaccine design. I was also involved in a large project in which we extended this model to detect regions of the HIV envelope glycoprotein that are targeted by broadly neutralising serum antibodies.

I have also worked on a collaborative project with researchers at the Beth Israel Deaconess Medical Centre, Harvard University in which a cohort of vaccinated and control monkeys were infected with the Simian Immunodeficiency Virus (SIV). We were interested in determining whether the viruses extracted from the vaccinated monkeys that became infected were genetically distinct from those of the control subjects; that is, did the vaccine successfully prevent optimal escape mutations and force the virus to explore different escape paths relative to the control group? A unique aspect of this so-called “sieve analysis” is that the SIV sequences of the infectious stock were known, which is typically not the case in studies of HIV evolution.

I am also interested in mathematical population genetics and have developed a diffusion process to estimate the selection coefficient of HIV mutants. While the use of diffusion processes to model allele frequency changes over time is not new to the field, the theory is grounded on the assumption that the selection coefficient is “small.” This is not likely to be the case for HIV escape mutations that rapidly rise to fixation in the intra-host viral population. I have developed a maximum likelihood procedure to efficiently estimate such “large” selection coefficients and their standard errors from longitudinal allele frequency data.

Publications

Conference Proceedings

  • Lacerda M, Haines LM and Fedderke JW. (2008). Purchasing power parity and uncovered interest parity in the presence of monetary and exchange rate regime shifts. South African Statistical Association Conference, Pretoria.
  • Lacerda M and Ardington C. (2005). Sequential regression multiple imputation for incomplete multivariate data using Markov chain Monte Carlo. South African Statistical Association Conference, Stellenbosch.