CompostBin: A DNA Composition-Based Algorithm for Binning Environmental Shotgun Reads

Chatterji, Sourav; Yamazaki, Ichitaro; Bai, Zhaojun; Eisen, Jonathan A.

doi:10.1007/978-3-540-78839-3_3

Cited by 117 publications

(107 citation statements)

References 39 publications

Supporting

Mentioning

107

Contrasting

Order By: Relevance

“…As most of the extant databases are highly biased in their representation of true diversity, such methods fail to find homologs for reads derived from novel species. On the other hand, composition-based methods rely on the intrinsic features of the reads such as oligomer/word distributions [15,3,5,13,12], codon usage preference [1] and GC composition [2] to ascertain the origin of the reads. The underlying basis is that the distribution of words in a DNA is specific to each species and undergoes only slight variations along the genome.…”

Section: Related Workmentioning

confidence: 99%

“…Most of the existing clustering methods are supervised and depend on the availability of reference data for training [15,3,19,5]. A metagenome may however, contain reads from unexplored phyla which cannot be labeled into one of the existing classes.…”

Section: Related Workmentioning

confidence: 99%

“…Most clustering methods until now have been relatively inaccurate in classifying short reads. The poor performance on short reads can be attributed to the high dimensionality and sparsity associated with the data [5].…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

A Two-Way Bayesian Mixture Model for Clustering in Metagenomics

Prabhakara

Acharya

2011

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. We present a new and efficient Bayesian mixture model based on Poisson and Multinomial distributions for clustering metagenomic reads by their species of origin. We use the relative abundance of different words along a genome to distinguish reads from different species. The distribution of word counts within a genome is accurately represented by a Poisson distribution. The Multinomial mixture model is derived as a standardized Poisson mixture model. The Bayesian network efficiently encodes the conditional dependencies between word counts in a DNA due to overlaps and hence is most consistent with the data. We present a two-way mixture model that captures the high dimensionality and sparsity associated with the data. Our method can cluster reads as short as 50 bps with accuracy over 80%. The Bayesian mixture models clearly outperform their Naive Bayes counterparts on datasets of varying abundances, divergences and read lengths. Our method attains comparable accuracy to that of state-of-art Scimm and converges at least 5 times faster than Scimm for all the cases tested. The reduced time taken, by our method, to obtain accurate results is highly significant and justifies the use of our proposed method to evaluate large metagenome datasets.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

A Two-Way Bayesian Mixture Model for Clustering in Metagenomics

Prabhakara

Acharya

2011

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…One category is the taxonomy-dependent methods (Brady and Salzberg, 2009;Haque et al, 2009;Huson et al, 2007;Krause et al, 2008;Matsen et al, 2010;McHardy et al, 2007;Meyer et al, 2008;Mohammed et al, 2011;Stark et al, 2010;Wood and Salzberg, 2014;Wu and Eisen, 2008), which compare reads with sequences in public databases or models inferred from public databases to group reads and determine which known species are present. The other category is the taxonomy-independent methods (Chatterji et al, 2008;Diaz et al, 2009;Wang et al, 2012Wang et al, , 2015Wu and Ye, 2011), which employ the difference of GC content (guanine-cytosine content), k-mer (a k base pairs long DNA segment) frequencies, etc., of different microbes in the same samples to bin reads.…”

mentioning

confidence: 99%

MBMC: An Effective Markov Chain Approach for Binning Metagenomic Reads from Environmental Shotgun Sequencing Projects

Wang

2016

OMICS: A Journal of Integrative Biology

View full text Add to dashboard Cite

Metagenomics is a next-generation omics field currently impacting postgenomic life sciences and medicine. Binning metagenomic reads is essential for the understanding of microbial function, compositions, and interactions in given environments. Despite the existence of dozens of computational methods for metagenomic read binning, it is still very challenging to bin reads. This is especially true for reads from unknown species, from species with similar abundance, and/or from low-abundance species in environmental samples. In this study, we developed a novel taxonomy-dependent and alignment-free approach called MBMC (Metagenomic Binning by Markov Chains). Different from all existing methods, MBMC bins reads by measuring the similarity of reads to the trained Markov chains for different taxa instead of directly comparing reads with known genomic sequences. By testing on more than 24 simulated and experimental datasets with species of similar abundance, species of low abundance, and/or unknown species, we report here that MBMC reliably grouped reads from different species into separate bins. Compared with four existing approaches, we demonstrated that the performance of MBMC was comparable with existing approaches when binning reads from sequenced species, and superior to existing approaches when binning reads from unknown species. MBMC is a pivotal tool for binning metagenomic reads in the current era of Big Data and postgenomic integrative biology. The MBMC software can be freely downloaded at http://hulab.ucf.edu/research/projects/ metagenomics/MBMC.html.

show abstract

“…frequency of occurrence for kmers found in broad classes of organisms (Teeling et al 2004, Chatterji et al 2008, and McHardy et al 2007). This approach is more scalable then sequence alignment but lacks the ability to provide detailed discrimination of the sample contents.…”

mentioning

confidence: 99%

FY11 Report on Metagenome Analysis using Pathogen Marker Libraries

Gardner

Allen

McLoughlin

et al. 2011

View full text Add to dashboard Cite

A method, sequence library, and software suite was invented to rapidly assess whether any member of a pre-specified list of threat organisms or their near neighbors is present in a metagenome. The system was designed to handle mega-to giga-bases of FASTA-formatted raw sequence reads from short or long read next generation sequencing platforms. The approach is to pre-calculate a viral and a bacterial -Pathogen Marker Library‖ (PML) containing sub-sequences specific to pathogens or their near neighbors. A list of expected matches comparing every bacterial or viral genome against the PML sequences is also pre-calculated. To analyze a metagenome, reads are compared to the PML, and observed PML-metagenome matches are compared to the expected PML-genome matches, and the ratio of observed relative to expected matches is reported. In other words, a 3-way comparison among the PML, metagenome, and existing genome sequences is used to quickly assess which (if any) species included in the PML is likely to be present in the metagenome, based on available sequence data.Our tests showed that the species with the most PML matches correctly indicated the organism sequenced for empirical metagenomes consisting of a cultured, relatively pure isolate. These runs completed in 1 minute to 3 hours on 12 CPU (1 thread/CPU), depending on the metagenome and PML. Using more threads on the same number of CPU resulted in speed improvements roughly proportional to the number of threads. Simulations indicated that detection sensitivity depends on both sequencing coverage levels for a species and the size of the PML: species were correctly detected even at ~0.003x coverage by the large PMLs, and at ~0.03x coverage by the smaller PMLs. Matches to true positive species were 3-4 orders of magnitude higher than to false positives. Simulations with short reads (36 nt and ~260 nt) showed that species were usually detected for metagenome coverage above 0.005x and coverage in the PML above 0.05x, and detection probability appears to be a function of both coverages. Multiple species could be detected simultaneously in a simulated low-coverage, complex metagenome, and the largest PML gave no false negative species and no false positive genera. The presence of multiple species was predicted in a complex metagenome from a human gut microbiome with 1.9 GB of short reads (75 nt); the species predicted were reasonable gut flora and no biothreat agents were detected, showing the feasibility of PML analysis of empirical complex metagenomes. BackgroundThe current approach to labeling metagenomic data is to use sequence alignment tools that align each read to a part of each reference sequence and report a summary of the top reference matches. The limitations of this approach are scalability and computational cost. Applying this approach to the Sargasso Sea metagenome data of 1.8 million reads (of length 700) would take approximately 42 days on a 64 CPU cluster (Huson et al. 2007). Using sequence alignment to label data sets produced by the newest high th...

show abstract

CompostBin: A DNA Composition-Based Algorithm for Binning Environmental Shotgun Reads

Cited by 117 publications

References 39 publications

A Two-Way Bayesian Mixture Model for Clustering in Metagenomics

A Two-Way Bayesian Mixture Model for Clustering in Metagenomics

MBMC: An Effective Markov Chain Approach for Binning Metagenomic Reads from Environmental Shotgun Sequencing Projects

FY11 Report on Metagenome Analysis using Pathogen Marker Libraries

Contact Info

Product

Resources

About