FastGroupII: a web-based bioinformatics platform for analyses of large 16S rDNA libraries

Abstract

Background: High-throughput sequencing makes it possible to rapidly obtain thousands of 16S rDNA sequences from environmental samples. Bioinformatic tools for the analyses of large 16S rDNA sequence databases are needed to comprehensively describe and compare these datasets.

Results: FastGroupII is a web-based bioinformatics platform to dereplicate large 16S rDNA libraries. FastGroupII provides users with the option of four different dereplication methods, performs rarefaction analysis, and automatically calculates the Shannon-Wiener Index and Chao1. FastGroupII was tested on a set of 16S rDNA sequences from coral-associated Bacteria. The different grouping algorithms produced similar, but not identical, results. This suggests that 16S rDNA datasets need to be analyzed in multiple ways when being used for community ecology studies.

Conclusion: FastGroupII is an effective bioinformatics tool for the trimming and dereplication of 16S rDNA sequences. Several standard diversity indices are calculated, and the raw sequences are prepared for downstream analyses.

Figure 1

Overview of the process for analyzing microbial communities using FastGroupII. A) Protocol for high-throughput sequencing of environmental microbial communities. B) Protocol for 16S rDNA analyses used in FastGroupII. Sequences are trimmed and dereplicated according to user-specified parameters. FastGroupII can perform rarefaction analysis, and calculate the Chao1 richness estimator and the Shannon-Wiener diversity index. The output from FastGroupII is formatted for submission to sequence classification programs such as BLAST [10] and RDP Classifier [11].

Figure 2

FastGroupII online analyses tool at FastGroupII Tools [6]. A FASTA formatted file containing the raw 16S rDNA sequences is first uploaded or pasted as the input file. The user then specifies the trimming and grouping criteria and selects the desired output. After submission, analysis is performed on the remote server and results are returned to the user on the same web page.

Figure 3

Rank-abundance curves predicted from the test dataset using FastGroupII and methods in DOTUR. The curves reveal similar grouping patterns predicted using the different methods. For clarity, the tails of singletons were excluded from the figure.