The VP7, VP4 and NSP4 genes of human rotavirus serotype 4 (G4) were analyzed to investigate intraserotypic variations. The techniques used included reverse transcription polymerase chain reaction with subtype specific primers, restriction fragment length polymorphism analysis and sequence analysis. Twelveisolates (nine from Japan and three from Kenya) and two standard strains (Hoehi, Odelia) were G4A P[8] Wa group NSP4. A standard strain (ST3) was G4A P[6] Wa group NSP4 and a strain (VA70) was G4B P[8] Wa group NSP4. These results show G4 rotaviruses can be divided into three combinations at the moment.Key words: Rotavirus, Serotype G4, VP7 gene, VP4 gene, NSP4 gene Rotavirus is a major etiologic agent that causes diarrhea in infants and children worldwide (3). In developing countries, rotavirus infection, coupled with malnutrition, leads to a severe and fatal outcome (1). At present, the genus Rotaviridae is classified into seven groups (groups A to G), based on group specific antigen. Group A rotavirus is the most prevalent (18,22). The genome of rotavirus comprises of 11 segments of double-stranded RNA, each of which encodes a distinct viral protein. Two outer capsid proteins, VP4 and VP7, are involved in virus neutralization. The neutralization specificities of VP4 and VP7 are designated as P and G serotypes for protease-sensitive protein and glycoprotein, respectively (5). Currently, two overlapping classification systems are used for VP4. P serotypes can be determined by P serotype specific monoclonal antibodies, however VP7 comprises of the greater percentage of the outer capsid protein and as such hyper-immune antisera raised to intact viral particles primarily recognize VP7 rather than VP4. For this reason, P typing is usually performed using reverse transcription polymerase chain reaction