Taxonomy annotation and guide tree errors in 16S rRNA databases.
ABSTRACT: Sequencing of the 16S ribosomal RNA (rRNA) gene is widely used to survey microbial communities. Specialized 16S rRNA databases have been developed to support this approach including Greengenes, RDP and SILVA. Most taxonomy annotations in these databases are predictions from sequence rather than authoritative assignments based on studies of type strains or isolates. In this work, I investigated the taxonomy annotations and guide trees provided by these databases. Using a blinded test, I estimated that the annotation error rate of the RDP database is ?10%. The branching orders of the Greengenes and SILVA guide trees were found to disagree at comparable rates with each other and with taxonomy annotations according to the training set (authoritative reference) provided by RDP, indicating that the trees have comparable quality. Pervasive conflicts between tree branching order and type strain taxonomies strongly suggest that the guide trees are unreliable guides to phylogeny. I found 249,490 identical sequences with conflicting annotations in SILVA v128 and Greengenes v13.5 at ranks up to phylum (7,804 conflicts), indicating that the annotation error rate in these databases is ?17%.
Project description:SILVA (from Latin silva, forest, http://www.arb-silva.de) is a comprehensive resource for up-to-date quality-controlled databases of aligned ribosomal RNA (rRNA) gene sequences from the Bacteria, Archaea and Eukaryota domains and supplementary online services. SILVA provides a manually curated taxonomy for all three domains of life, based on representative phylogenetic trees for the small- and large-subunit rRNA genes. This article describes the improvements the SILVA taxonomy has undergone in the last 3 years. Specifically we are focusing on the curation process, the various resources used for curation and the comparison of the SILVA taxonomy with Greengenes and RDP-II taxonomies. Our comparisons not only revealed a reasonable overlap between the taxa names, but also points to significant differences in both names and numbers of taxa between the three resources.
Project description:A key step in microbiome sequencing analysis is read assignment to taxonomic units. This is often performed using one of four taxonomic classifications, namely SILVA, RDP, Greengenes or NCBI. It is unclear how similar these are and how to compare analysis results that are based on different taxonomies.We provide a method and software for mapping taxonomic entities from one taxonomy onto another. We use it to compare the four taxonomies and the Open Tree of life Taxonomy (OTT).While we find that SILVA, RDP and Greengenes map well into NCBI, and all four map well into the OTT, mapping the two larger taxonomies on to the smaller ones is problematic.
Project description:Molecular sequences in public databases are mostly annotated by the submitting authors without further validation. This procedure can generate erroneous taxonomic sequence labels. Mislabeled sequences are hard to identify, and they can induce downstream errors because new sequences are typically annotated using existing ones. Furthermore, taxonomic mislabelings in reference sequence databases can bias metagenetic studies which rely on the taxonomy. Despite significant efforts to improve the quality of taxonomic annotations, the curation rate is low because of the labor-intensive manual curation process. Here, we present SATIVA, a phylogeny-aware method to automatically identify taxonomically mislabeled sequences ('mislabels') using statistical models of evolution. We use the Evolutionary Placement Algorithm (EPA) to detect and score sequences whose taxonomic annotation is not supported by the underlying phylogenetic signal, and automatically propose a corrected taxonomic classification for those. Using simulated data, we show that our method attains high accuracy for identification (96.9% sensitivity/91.7% precision) as well as correction (94.9% sensitivity/89.9% precision) of mislabels. Furthermore, an analysis of four widely used microbial 16S reference databases (Greengenes, LTP, RDP and SILVA) indicates that they currently contain between 0.2% and 2.5% mislabels. Finally, we use SATIVA to perform an in-depth evaluation of alternative taxonomies for Cyanobacteria. SATIVA is freely available at https://github.com/amkozlov/sativa.
Project description:There is currently no criterion to select appropriate bioinformatics tools and reference databases for analysis of 16S rRNA amplicon data in the human oral microbiome. Our study aims to determine the influence of multiple tools and reference databases on ?-diversity measurements and ?-diversity comparisons analyzing the human oral microbiome. We compared the results of taxonomical classification by Greengenes, the Human Oral Microbiome Database (HOMD), National Center for Biotechnology Information (NCBI) 16S, SILVA, and the Ribosomal Database Project (RDP) using Quantitative Insights Into Microbial Ecology (QIIME) and the Divisive Amplicon Denoising Algorithm (DADA2). There were 15 phyla present in all of the analyses, four phyla exclusive to certain databases, and different numbers of genera were identified in each database. Common genera found in the oral microbiome, such as Veillonella, Rothia, and Prevotella, are annotated by all databases; however, less common genera, such as Bulleidia and Paludibacter, are only annotated by large databases, such as Greengenes. Our results indicate that using different reference databases in 16S rRNA amplicon data analysis could lead to different taxonomic compositions, especially at genus level. There are a variety of databases available, but there are no defined criteria for data curation and validation of annotations, which can affect the accuracy and reproducibility of results, making it difficult to compare data across studies.
Project description:SUMMARY:We developed the metagenomeFeatures R Bioconductor package along with annotation packages for three 16S rRNA databases (Greengenes, RDP and SILVA) to facilitate working with 16S rRNA databases and marker-gene survey feature data. The metagenomeFeatures package defines two classes, MgDb for working with 16S rRNA sequence databases, and mgFeatures for marker-gene survey feature data. The associated annotation packages provide a consistent interface to the different databases facilitating database comparison and exploration. The mgFeatures-class represents a crucial step in the development of a common data structure for working with 16S marker-gene survey data in R. AVAILABILITY AND IMPLEMENTATION:https://bioconductor.org/packages/release/bioc/html/metagenomeFeatures.html. SUPPLEMENTARY INFORMATION:Supplementary material is available at Bioinformatics online.
Project description:Taxonomy identification is fundamental to all microbiology studies. Particularly in metagenomics, which identify the composition of microorganisms using thousands of sequences, its importance is even greater. Identification is inevitably affected by the choice of database. This study was conducted to evaluate the accuracy of three widely used 16S databases, Greengenes, Silva, and EzBioCloud, and to suggest basic guidelines for selecting reference databases. Using public mock community data, each database was used to assign taxonomy and to test its accuracy. We showed that EzBioCloud performs well compared to other existing databases.
Project description:BACKGROUND: Microbial ecologists now routinely utilize next-generation sequencing methods to assess microbial diversity in the environment. One tool heavily utilized by many groups is the Naïve Bayesian Classifier developed by the Ribosomal Database Project (RDP-NBC). However, the consistency and confidence of classifications provided by the RDP-NBC is dependent on the training set utilized. RESULTS: We explored the stability of classification of honey bee gut microbiota sequences by the RDP-NBC utilizing three publically available ribosomal RNA sequence databases as training sets: ARB-SILVA, Greengenes and RDP. We found that the inclusion of previously published, high-quality, full-length sequences from 16S rRNA clone libraries improved the precision in classification of novel bee-associated sequences. Specifically, by including bee-specific 16S rRNA gene sequences a larger fraction of sequences were classified at a higher confidence by the RDP-NBC (based on bootstrap scores). CONCLUSIONS: Results from the analysis of these bee-associated sequences have ramifications for other environments represented by few sequences in the public databases or few bacterial isolates. We conclude that for the exploration of relatively novel habitats, the inclusion of high-quality, full-length 16S rRNA gene sequences allows for a more confident taxonomic classification.
Project description:Sequencing of taxonomic or phylogenetic markers is becoming a fast and efficient method for studying environmental microbial communities. This has resulted in a steadily growing collection of marker sequences, most notably of the small-subunit (SSU) ribosomal RNA gene, and an increased understanding of microbial phylogeny, diversity and community composition patterns. However, to utilize these large datasets together with new sequencing technologies, a reliable and flexible system for taxonomic classification is critical. We developed CREST (Classification Resources for Environmental Sequence Tags), a set of resources and tools for generating and utilizing custom taxonomies and reference datasets for classification of environmental sequences. CREST uses an alignment-based classification method with the lowest common ancestor algorithm. It also uses explicit rank similarity criteria to reduce false positives and identify novel taxa. We implemented this method in a web server, a command line tool and the graphical user interfaced program MEGAN. Further, we provide the SSU rRNA reference database and taxonomy SilvaMod, derived from the publicly available SILVA SSURef, for classification of sequences from bacteria, archaea and eukaryotes. Using cross-validation and environmental datasets, we compared the performance of CREST and SilvaMod to the RDP Classifier. We also utilized Greengenes as a reference database, both with CREST and the RDP Classifier. These analyses indicate that CREST performs better than alignment-free methods with higher recall rate (sensitivity) as well as precision, and with the ability to accurately identify most sequences from novel taxa. Classification using SilvaMod performed better than with Greengenes, particularly when applied to environmental sequences. CREST is freely available under a GNU General Public License (v3) from http://apps.cbu.uib.no/crest and http://lcaclassifier.googlecode.com.
Project description:Next-generation sequencing has provided powerful tools to conduct microbial ecology studies. Analysis of community composition relies on annotated databases of curated sequences to provide taxonomic assignments; however, these databases occasionally have errors with implications for downstream analyses. Systemic taxonomic errors were discovered in Greengenes database (v13_5 and 13_8) related to orders Vibrionales and Alteromonadales. These orders have family level annotations that were erroneous at least one taxonomic level, e.g., 100% of sequences assigned to the Pseudoalteromonadaceae family were placed improperly in Vibrionales (rather than Alteromonadales) and >20% of these sequences were indeed Vibrio spp. but were improperly assigned to the Pseudoalteromonadaceae family (rather than to Vibrionaceae). Use of this database is common; we identified 68 peer-reviewed papers since 2013 that likely included erroneous annotations specifically associated with Vibrionales and Pseudoalteromonadaceae, with 20 explicitly stating the incorrect taxonomy. Erroneous assignments using these specific versions of Greengenes can lead to incorrect conclusions, especially in marine systems where these taxa are commonly encountered as conditionally rare organisms and potential pathogens.
Project description:Methane is formed by methanogenic archaea in the rumen as one of the end products of feed fermentation in the ruminant digestive tract. To develop strategies to mitigate anthropogenic methane emissions due to ruminant farming, and to understand rumen microbial differences in animal feed conversion efficiency, it is essential that methanogens can be identified and taxonomically classified with high accuracy. Currently available taxonomic frameworks offer only limited resolution beyond the genus level for taxonomic assignments of sequence data stemming from high throughput sequencing technologies. Therefore, we have developed a QIIME-compatible database (DB) designed for species-level taxonomic assignment of 16S rRNA gene amplicon data targeting methanogenic archaea from the rumen, and from animal and human intestinal tracts. Called RIM-DB (Rumen and Intestinal Methanogen-DB), it contains a set of 2,379 almost full-length chimera-checked 16S rRNA gene sequences, including 20 previously unpublished sequences from isolates from three different orders. The taxonomy encompasses the recently-proposed seventh order of methanogens, the Methanomassiliicoccales, and allows differentiation between defined groups within this order. Sequence reads from rumen contents from a range of ruminant-diet combinations were taxonomically assigned using RIM-DB, Greengenes and SILVA. This comparison clearly showed that taxonomic assignments with RIM-DB resulted in the most detailed assignment, and only RIM-DB taxonomic assignments allowed methanogens to be distinguished taxonomically at the species level. RIM-DB complements the use of comprehensive databases such as Greengenes and SILVA for community structure analysis of methanogens from the rumen and other intestinal environments, and allows identification of target species for methane mitigation strategies.