Streptococcus pneumoniae (the pneumococcus) is a leading cause of morbidity and mortality worldwide, causing otitis media, pneumonia, sepsis and meningitis in children, the elderly and the immunocompromised. Over 90 immunologically-distinct serotypes of pneumococcus have been reported, which vary in the biochemical structure of their capsule, encoded by the capsular polysaccharide (cps) locus. Accurate serotyping of pneumococci is required to track changes in serotype prevalence following vaccine introduction. Reference cps DNA sequences for the 90+ serotypes used for molecular serotyping are derived primarily from high-income countries. Data on serotype prevalence and distribution from high-income countries were a key driver in decisions surrounding the serotypes to be included in vaccines. These vaccines have subsequently been introduced into low and middle-income countries (LMICs), where limited data on pneumococcal population structure are often available. Our group has used DNA microarray to monitor serotype prevalence in LMICs in the Asia-Pacific region including in Fiji, Mongolia, Vietnam, Indonesia and Lao PDR. We identified two novel cps loci (designated 11F-like and 33F-like) in nasopharyngeal swabs both from healthy children and children with pneumonia. The genetic variation in the 11F-like cps locus results in changes to the antigenic properties of the capsule such that it does not encodes the production of an 11F capsule. This results in discrepant serotyping calls depending on the serotyping method used. Although the serological properties of the 33F-like capsule match that of the traditional serotype 33F, the 33F-like cps locus appears to be a mosaic of genes likely derived from other pneumococcal serotypes. This includes the presence of a wcyO acetyltransferase pseudogene, which has not been reported in serotype 33F previously. This work has important implications for improving the accuracy of serotyping methods used for measuring vaccine impact and invasive pneumococcal disease surveillance, particularly in LMICs, where pneumococcal serotype diversity is poorly understood.