A fuzzy classifier to taxonomically group DNA fragments within a metagenome

Nasser, Sara; Breland, Adrienne; Harris, Frederick C.; Nicolescu, Mircea

doi:10.1109/nafips.2008.4531252

Cited by 6 publications

(9 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The distillation of complex datasets into defined groupings facilitates data interpretation and new hypothesis development based on average group properties and composition. Similar clustering approaches, including the fuzzy–c-means algorithm, have been applied in the field of environmental metagenomics, enabling the binning of sequence reads within complex data sets (e.g., Nasser et al, 2008 ; Liu et al, 2015 ; Lu et al, 2016 ).…”

Section: Discussionmentioning

confidence: 99%

Stable Isotope Phenotyping via Cluster Analysis of NanoSIMS Data As a Method for Characterizing Distinct Microbial Ecophysiologies and Sulfur-Cycling in the Environment

et al. 2016

View full text Add to dashboard Cite

Stable isotope probing (SIP) is a valuable tool for gaining insights into ecophysiology and biogeochemical cycling of environmental microbial communities by tracking isotopically labeled compounds into cellular macromolecules as well as into byproducts of respiration. SIP, in conjunction with nanoscale secondary ion mass spectrometry (NanoSIMS), allows for the visualization of isotope incorporation at the single cell level. In this manner, both active cells within a diverse population as well as heterogeneity in metabolism within a homogeneous population can be observed. The ecophysiological implications of these single cell stable isotope measurements are often limited to the taxonomic resolution of paired fluorescence in situ hybridization (FISH) microscopy. Here we introduce a taxonomy-independent method using multi-isotope SIP and NanoSIMS for identifying and grouping phenotypically similar microbial cells by their chemical and isotopic fingerprint. This method was applied to SIP experiments in a sulfur-cycling biofilm collected from sulfidic intertidal vents amended with 13C-acetate, 15N-ammonium, and 33S-sulfate. Using a cluster analysis technique based on fuzzy c-means to group cells according to their isotope (13C/12C, 15N/14N, and 33S/32S) and elemental ratio (C/CN and S/CN) profiles, our analysis partitioned ~2200 cellular regions of interest (ROIs) into five distinct groups. These isotope phenotype groupings are reflective of the variation in labeled substrate uptake by cells in a multispecies metabolic network dominated by Gamma- and Deltaproteobacteria. Populations independently grouped by isotope phenotype were subsequently compared with paired FISH data, demonstrating a single coherent deltaproteobacterial cluster and multiple gammaproteobacterial groups, highlighting the distinct ecophysiologies of spatially-associated microbes within the sulfur-cycling biofilm from White Point Beach, CA.

show abstract

Section: Discussionmentioning

confidence: 99%

Stable Isotope Phenotyping via Cluster Analysis of NanoSIMS Data As a Method for Characterizing Distinct Microbial Ecophysiologies and Sulfur-Cycling in the Environment

et al. 2016

View full text Add to dashboard Cite

show abstract

“…Literature abounds in methods for classifying (as opposed to clustering) metagenome reads into taxon-specific bins [22,4,23]. Current approaches to metagenomics clustering can be classified into two main categories: similaritybased and composition-based methods.…”

Section: Related Workmentioning

confidence: 99%

“…Its accuracy of assignment drops drastically (to just 7.1% at genus level) for short reads and reads from unknown species. Nasser et al [23] demonstrated that a k-means based fuzzy classifier, trained using a maximal order Markov chain, can separate fragments that are about 1 kbp long at the phylum level with a high accuracy. Rosen et al trained a Naive Bayes classifier using publicly available microbial genomes [28].…”

Section: Related Workmentioning

confidence: 99%

“…Clustering provides deeper insight into the structure of the community. It can lead to faster and more robust assembly by reducing the search space [23].…”

Section: Introductionmentioning

confidence: 99%

“…Most of the existing clustering methods are supervised and depend on the availability of reference data for training [14,22,4,23,7,28] . A metagenomic dataset, may however, contain reads from unexplored phyla which cannot be labeled into one of the existing classes.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A two-way multi-dimensional mixture model for clustering metagenomic sequences

Prabhakara

Acharya

2011

Proceedings of the 2nd ACM Conference on Bioinformatics, Computational Biology and Biomedicine

View full text Add to dashboard Cite

Motivation:A major challenge facing metagenomics is the development of tools for the characterization of functional and taxonomic content of vast amounts of short metagenome reads. The efficacy of clustering methods depends on the number of reads in the dataset, the read length and relative abundances of source genomes in the microbial community.Results: In this paper, we formulate an unsupervised naive Bayes multi-species, multi-dimensional mixture model for reads from a metagenome. We use the proposed model to cluster metagenomic reads by their species of origin and to characterize the abundance of each species. We model the distribution of word counts along a genome as a Gaussian for shorter, frequent words and as a Poisson for longer words that are rare. We employ either a mixture of Gaussians or mixture of Poissons to model reads within each bin. An additional reason to use these distributions is their flexibility and ease of parameter estimation. Such a paradigm characterizes the compositional heterogeneity of the words along a genome, signifying its genome signature. Further, we handle the high-dimensionality and sparsity associated with the data, by grouping the set of words comprising the reads, resulting in a two-way mixture model. Finally, we derive an unsupervised Expectation Maximization algorithm for the models. Our method provides a general statistical framework for modeling metagenome reads. We demonstrate the accuracy and applicability of this method on simulated and real metagenomes. Our method can accurately cluster reads as short as 100 bps and estimate the species abundance as well. Our method outperforms LikelyBin, another unsupervised composition-based binning method for metagenomes, on datasets of varying abundances, divergences and read lengths.

show abstract

SIMCOMP: A Hybrid Soft Clustering of Metagenome Reads

Prabhakara

Acharya

2010

Pattern Recognition in Bioinformatics

View full text Add to dashboard Cite

A fuzzy classifier to taxonomically group DNA fragments within a metagenome

Cited by 6 publications

References 27 publications

Stable Isotope Phenotyping via Cluster Analysis of NanoSIMS Data As a Method for Characterizing Distinct Microbial Ecophysiologies and Sulfur-Cycling in the Environment

Stable Isotope Phenotyping via Cluster Analysis of NanoSIMS Data As a Method for Characterizing Distinct Microbial Ecophysiologies and Sulfur-Cycling in the Environment

A two-way multi-dimensional mixture model for clustering metagenomic sequences

SIMCOMP: A Hybrid Soft Clustering of Metagenome Reads

Contact Info

Product

Resources

About