Abstract
Streptococcus pneumoniae is one of the world's leading bacterial pathogens, responsible for pneumonia, septicaemia and meningitis. Asymptomatic colonisation of the nasopharynx is considered to be a prerequisite for these severe infections, however little is understood about the biological changes that permit the pneumococcus to switch from asymptomatic coloniser to invasive pathogen. A phase variable type I restriction-modification (R-M) system (SpnIII) has been linked to a change in capsule expression and to the ability to successfully colonise the murine nasopharynx. Using our laboratory data, we have developed a Markov change model that allows prediction of the expected level of phase variation within a population, and as a result measures when populations deviate from those expected at random. Using this model, we have analysed samples from the Experimental Human Pneumococcal Carriage (EHPC) project. Here we show, through mathematical modelling, that the patterns of dominant SpnIII alleles expressed in the human nasopharynx are significantly different than those predicted by stochastic switching alone. Our inter-disciplinary work demonstrates that the expression of alternative methylation patterns should be an important consideration in studies of pneumococcal colonisation.