The oral cavity is an important window for microbial communication between the environment and the human body. The oral microbiome plays an important role in human health. However, compared to the gut microbiome, the oral microbiome has been poorly explored. Here, we analyzed 404 datasets from human oral saliva samples published by the Earth Microbiome Project (EMP) and compared them with 815 samples from the human gut, nose/pharynx, and skin. The diversity of the human saliva microbiome varied significantly among individuals, and the community compositions were complex and diverse. The saliva microbiome showed the lowest species diversity among the four environment types. Human oral habitats shared a small core bacterial community containing only 14 operational taxonomic units (OTUs) under 5 phyla, which occupied over 75% of the sequence abundance. For the four habitats, the core taxa of the saliva microbiome had the greatest impact on saliva habitats than other habitats and were mostly unique. In addition, the saliva microbiome showed significant differences in the populations of different regions, which may be determined by the living environment and lifestyle/dietary habits. Finally, the correlation analysis showed high similarity between the saliva microbiome and the microbiomes of Aerosol (non-saline) and Surface (non-saline), i.e., two environment types closely related to human, suggesting that contact and shared environment being the driving factors of microbial transmission. Together, these findings expand our understanding of human oral diversity and biogeography.

Quantitative evaluation of protein structural evolution is important for our understanding of protein biological functions and their evolutionary adaptation, and is useful in guiding protein engineering. However, compared to the models for sequence evolution, the quantitative models for protein structural evolution received less attention. Ancient protein superfamilies are often considered versatile, allowing genetic and functional diversifications during long-term evolution. In this study, we investigated the quantitative impacts of sequence variations on the structural evolution of homologues in 68 ancient protein superfamilies that exist widely in sequenced eukaryotic, bacterial and archaeal genomes. We found that the accumulated structural variations within ancient superfamilies could be explained largely by a bilinear model that simultaneously considers amino acid substitution and insertion/deletion (indel). Both substitutions and indels are essential for explaining the structural variations within ancient superfamilies. For those ancient superfamilies with high bilinear multiple correlation coefficients, the influence of each unit of substitution or indel on structural variations is almost constant within each superfamily, but varies greatly among different superfamilies. The influence of each unit indel on structural variations is always larger than that of each unit substitution within each superfamily, but the accumulated contributions of indels to structural variations are lower than those of substitutions in most superfamilies. The total contributions of sequence indels and substitutions (46% and 54%, respectively) to the structural variations that result from sequence variations are slightly different in ancient superfamilies. Structural variations within ancient protein superfamilies accumulated under the significantly bilinear influence of amino acid substitutions and indels in sequences. Both substitutions and indels are essential for explaining the structural variations within ancient superfamilies. For those structural variations resulting from sequence variations, the total contribution of indels is slightly lower than that of amino acid substitutions. The regular clock exists not only in protein sequences, but also probably in protein structures.