Daniel R. Mende; Alison S. Waller; Shinichi Sunagawa; Aino I. Järvelin; Michelle M. Chan; Manimozhiyan Arumugam; Jeroen Raes; Peer Bork

“ All of the sequencing technologies perform comparably when the coverage per organism is relatively high as in the 10 genomes metagenome. But for the more complex communities (100 genomes) Illumina performs better due to the greater sequencing depth achieved. However, for the metagenomes with 400 genomes even the sequencing depth achieved with Illumina does not make up for the lower coverage of each genome and Sanger performs well due to the long read lengths. ”

“ Our results show the importance of good quality control as the Illumina assemblies greatly improved after rigorous quality filtering and trimming.
After this initial step there are a number of analyses that one can perform, but to decrease the complexity of the data, the quality controlled reads are usually assembled into contigs. Our comparison of assemblies from different sequencing technologies reveals that each assembly has different characteristics depending on community composition and sequencing technology. ”

“ Although there was no assembly of Sanger reads into contigs, the Sanger reads still represent the best ‘assembly’ with 10 times more fragments over 500 bp, than the Illumina assembly and over 10 times more genes with functional annotation. These differences can be attributed to the differences in read length and especially sequencing depth, but both parameters are intrinsic to the different sequencing technologies., as the cost of the sequencing technology is directly related to the sequencing depth. ”

“ The COG abundance profiles were compared to each other using Jensen-Shannon divergence and the distance matrix was then analyzed plotted using PCoA. The COG abundance profiles were analyzed plotted using PCA. The dots are colored by sequencing method: Illumina contigs (red) , Illumina scaftigs (light blue) , Sanger (yellow) and pyrosequencing (blue) . Discussion
The first step in any metagenomic data analysis should be raw data treatment. This includes quality control and removal of contamination (eg. human contamination in metagenomic studies of the human gut) . ”

“ The protein-coding genes are then used to predict the functional repertoire encoded in the metagenomes and the phylogenetic composition can be estimated using a variety of methods [18] . Data analysis pipeline tools like SmashCommunity [19] , MG-RAST [20] , IMG/M [21] and Metarep [22] , are complemented by numerous special purpose tools, and they all need to be validated. ”

“ The authors have declared that no competing interests exist.Funding: This research received funding from MetaHIT, the Natural Sciences and Engineering Research Council of Canada (funding A.S.W.) and European Molecular Biology Laboratory. ”

“ JCVI Metagenomics Reports - an open source tool for high-performance comparative metagenomics.”, Bioinformatics (Oxford, England) 26: 2631–2632 1 2 3 4 5 6 Mavromatis, K; Ivanova, N; Barry, K; Shapiro, H & Goltsman, E, “Use of simulated data sets to evaluate the fidelity of metagenomic processing methods.”, Nature methods 4: 495–500, pmid:17468765 1 2 Richter, DC; Ott, F; Auch, AF; Schmid, R & Huson, DH, “MetaSim: a sequencing simulator for genomics and metagenomics.”, PloS one 3: e3373–e3373, pmid:18841204 ↑ Angly, FE ”