An official website of the United States government
Taxonomies are fundamental to many real-world applications in various domains, serving as structural representations of knowledge. To deal with the increasing volume of new concepts needed to be organized as taxonomies, researchers turn to automatically completion of an existing taxonomy with new concepts. In this paper, we propose TaxoEnrich, a new taxonomy completion framework, which effectively leverages both semantic features and structural information in the existing taxonomy and offers a better representation of candidate position to boost the performance of taxonomy completion. Specifically, TaxoEnrich consists of four components: (1) taxonomy-contextualized embedding which incorporates both semantic meanings of concept and taxonomic relations based on powerful pretrained language models; (2) a taxonomy-aware sequential encoder which learns candidate position representations by encoding the structural information of taxonomy; (3) a query-aware sibling encoder which adaptively aggregates candidate siblings to augment candidate position representations based on their importance to the query-position matching; (4) a query-position matching model which extends existing work with our new candidate position representations. Extensive experiments on four large real-world datasets from different domains show that TaxoEnrich achieves the best performance among all evaluation metrics and outperforms previous state-of-the-art methods by a large margin.
more »
« less
LINgroups as a principled approach to compare and integrate multiple bacterial taxonomies
Traditional taxonomy provides a hierarchical organization of bacte- ria and archaea across taxonomic ranks from kingdom to subspecies. More recently, bacterial taxonomy has been more robustly quanti- fied using comparisons of sequenced genomes, as in the Genome Taxonomy Database (GTDB), resolving down to genera and species. Such taxonomies have proven useful in many contexts, yet lack the flexibility and resolution of a more fine-grained approach. We apply our Life Identification Number (LIN) approach as a com- mon, quantitative framework to tie existing (and future) bacterial taxonomies together, increase the resolution of genome-based dis- crimination of taxa, and extend taxonomic identification below the species level in a principled way. We utilize our existing concept of a LINgroup as an organizational concept for microorganisms that are closely related by overall genomic similarity, to help resolve some of the confusions and unforeseen negative effects of nomen- clature changes of microbes due to genome-based reclassification. Our results obtained from experimentation demonstrate the value of LINs and LINgroups in mapping between taxonomies, translat- ing between different nomenclatures, and integrating them into a single taxonomic framework.
more »
« less
- Award ID(s):
- 2018522
- PAR ID:
- 10347236
- Date Published:
- Journal Name:
- Proceedings of the 13th ACM International Conference on Bioinformatics, Computational Biology and Health Informatics
- Page Range / eLocation ID:
- 1 to 7
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
ABSTRACT Genus assignment is fundamental in the characterization of microbes, yet there is currently no unambiguous way to demarcate genera solely using standard genomic relatedness indices. Here, we propose an approach to demarcate genera that relies on the combined use of the average nucleotide identity, genome alignment fraction, and the distinction between type- and non-type species. More than 3,500 genomes representing type strains of species from >850 genera of either bacterial or archaeal lineages were tested. Over 140 genera were analyzed in detail within the taxonomic context of order/family. Significant genomic differences between members of a genus and type species of other genera in the same order/family were conserved in 94% of the cases. Nearly 90% (92% if polyphyletic genera are excluded) of the type strains were classified in agreement with current taxonomy. The 448 type strains that need reclassification directly impact 33% of the genera analyzed in detail. The results provide a first line of evidence that the combination of genomic indices provides added resolution to effectively demarcate genera within the taxonomic framework that is currently based on the 16S rRNA gene. We also identify the emergence of natural breakpoints at the genome level that can further help in the circumscription of taxa, increasing the proportion of directly impacted genera to at least 43% and pointing at inaccuracies on the use of the 16S rRNA gene as a taxonomic marker, despite its precision. Altogether, these results suggest that genomic coherence is an emergent property of genera in Bacteria and Archaea . IMPORTANCE In recent decades, the taxonomy of Bacteria and Archaea , and therefore genus designation, has been largely based on the use of a single ribosomal gene, the 16S rRNA gene, as a taxonomic marker. We propose an approach to delineate genera that excludes the direct use of the 16S rRNA gene and focuses on a standard genome relatedness index, the average nucleotide identity. Our findings are of importance to the microbiology community because the emergent properties of Bacteria and Archaea that are identified in this study will help assign genera with higher taxonomic resolution.more » « less
-
Taxonomies consist of machine-interpretable semantics and provide valuable knowledge for many web applications. For example, online retailers (e.g., Amazon and eBay) use taxonomies for product recommendation, and web search engines (e.g., Google and Bing) leverage taxonomies to enhance query understanding. Enormous efforts have been made on constructing taxonomies eithermanually or semi-automatically. However, with the fast-growing volume of web content, existing taxonomies will become outdated and fail to capture emerging knowledge. Therefore, in many applications, dynamic expansions of an existing taxonomy are in great demand. In this paper, we study how to expand an existing taxonomy by adding a set of new concepts. We propose a novel self-supervised framework, named TaxoExpan, which automatically generates a set of ⟨query concept, anchor concept⟩ pairs from the existing taxonomy as training data. Using such self-supervision data, TaxoExpan learns a model to predict whether a query concept is the direct hyponym of an anchor concept. We develop two innovative techniques in TaxoExpan: (1) a position-enhanced graph neural network that encodes the local structure of an anchor concept in the existing taxonomy, and (2) a noise-robust training objective that enables the learned model to be insensitive to the label noise in the self-supervision data. Extensive experiments on three large-scale datasets from different domains demonstrate both the effectiveness and the efficiency of TaxoExpan for taxonomy expansion.more » « less
-
Genomics has put prokaryotic rank-based taxonomy on a solid phylogenetic foundation. However, most taxonomic ranks were set long before the advent of DNA sequencing and genomics. In this concept paper, we thus ask the following question: should prokaryotic classification schemes besides the current phylum-to-species ranks be explored, developed, and incorporated into scientific discourse? Could such alternative schemes provide better solutions to the basic need of science and society for which taxonomy was developed, namely, precise and meaningful identification? A neutral genome-similarity based framework is then described that could allow alternative classification schemes to be explored, compared, and translated into each other without having to choose only one as the gold standard. Classification schemes could thus continue to evolve and be selected according to their benefits and based on how well they fulfill the need for prokaryotic identification.more » « less
-
Taxonomies serve many applications with a structural representation of knowledge. To incorporate emerging concepts into existing taxonomies, the task of taxonomy completion aims to find suitable positions for emerging query concepts. Previous work captured homogeneous token-level interactions inside a concatenation of the query concept term and definition using pre-trained language mod- els. However, they ignored the token-level interactions between the term and definition of the query concepts and their related concepts. In this work, we propose to capture heterogeneous token-level interactions between the different textual components of concepts that have different types of relations. We design a relation-aware mutual attention module (RAMA) to learn such interactions for taxonomy completion. Experimental results demonstrate that our new taxonomy completion framework based on RAMA achieves the state-of-the-art performance on six taxonomy datasets.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1145/3535508.3545546
