Small size sampling of large size metagenomic sequences and its environmental applications

다국어 초록 (Multilingual Abstract)

Metagenomics is regarded as one of the great opportunities of environmental biotechnology, as high throughput DNA sequencing is getting popular. However, it requires a large amount of computing power and heavy time consumption for analyzing huge size of DNA sequence data, which is main barrier to the applicability of metagenomics in environmental applications. To shorten computation time with less computing resources, a simple concept of rarefaction technique was utilized in this study. The research objectives of this study were to examine the accuracy of the taxonomical and functional analyses by the small sampling method optimized in this work, and to provide specific guideline information on the size and the way of small sampling, and the levels of taxonomical and functional information.
As the first step, metagenomic sequences of a mock community, which was intentionally made of known bacterial strains to get a mixed genome, were examined with BLAST search. In total 792 samples were made with 6 different sampling methods. Since the real taxonomic information of the mock community is known, one can evaluate how well a sample represents the original taxonomic information. As the second step, the sampling method was also tested for 10 known results of previously reported metagenome researches, comparing taxonomic information, as well as gene function information.
The BLAST search test showed that the increase of the sample size results the increase of accuracy in taxonomic distribution, following the statistical tendency. Assuming a general normal distribution model, a sample with 5,000 reads was suggested as an option with 1% margin of error and 85% confidence. Meanwhile, among the 6 sampling methods, the convenience sampling method of selecting from the end showed the noticeably worst representativeness. While other method showed much better results than that, the systematic sampling of uniform selection showed the best result.
Among the results from previously reported metagenome sequence results, 9 out of 10 cases of both the most annotated phyla and the most annotated classes were matched between the samples and the originals. The Shannon diversity indexes calculated from them also showed the similarity between ones from the samples and ones from the originals in their tendency. Additionally, COG and SEED Subsystems function types annotated by MG-RAST showed the similarity. From the findings, a systematic method for determining size of credible sample was proposed with its users’ guideline including a flow chart for designing the small sampling method and evaluating its errors.
Since this sampling method only costs short time and small computing resources, one can use this approach to develop a standard or a protocol to preview or pre-check metagenomics data, before performing more accurate analysis with original full sequences. This will make metagenomics to be more efficient, enabling its environmental applications in wider area.

참고문헌 (Reference)

1. 농업용수 수질개선사업, 어디까지 왔나? PRI Focus 20, Kim, H., ISSN:2288-0682, , 2013