Up to Species-level Community Analysis of Human Gut Microbiota by 16S rRNA Amplicon Pyrosequencing

Abstract

Pyrosequencing-based 16S rRNA profiling has become a common powerful tool to obtain the community structure of gastrointestinal tract microbiota, but it is still hard to process the massive amount of sequence data into microbial composition data, especially at the species level. Here we propose a new approach in combining the quantitative insights into microbial ecology (QIIME), Mothur and ribosomal database project (RDP) programs to efficiently process 454 pyrosequence data to bacterial composition data up to the species level. It was demonstrated to precisely convert batch sequence data of 16S rRNA V6-V8 amplicons obtained from adult Singaporean fecal samples to taxonomically annotated biota data.

Keywords: 16S rRNA gene; human gut microbiota; pyrosequence.

Fig. 1.

Flowchart of computational processing of 454 pyrosequence data to individual microbiota data.

Fig. 2.

Result of a chimera check of 15,578 nonredundant sequences. The scores of each sequence in de novo Uchime and DB Uchime were two-dimensionally plotted. Grey dots represent sequences with read counts of less than 10. Black dots represent sequences with the read counts of more than 9, and sequences with more than 99 counts are surrounded by a square. Vertical and horizontal dotted lines show the cut off threshold (score = 0.6) used for the chimera detection except in the case of SG.9_C1697.

Fig. 3.

Population distribution of 37 (A), 19 (B) and 4 (C) common genera, families and phyla, respectively, among the 28 Singaporean subjects. The relative abundance of each taxonomic group was calculated by dividing the read counts of identified sequences by the individual`s total read number. The 37 genera, 19 families, and 4 phyla were selected as they were detected in more than a half of our Singaporean subjects.

Fig. 4.

Comparison of the relative abundances of Bacteroides (A) and Bifidobacterium (B) determined by 16S rRNA amplicon pyrosequencing with those determined by quantitative real-time PCR. The relative abundance in the pyrosequencing data was calculated by dividing the number of reads identified to genus Bacteroides or Bifidobacterium by the total read counts in each subject. In the quantitative PCR, group-specific primers targeting the Bacteroides fragilis group and genus Bifidobacterium were used, respectively [5].