Phymm and PhymmBL: Metagenomic Phylogenetic Classification with Interpolated Markov Models

By Arthur Brady, Steven L. Salzberg. Nature Methods, 2009

511 citations recorded by the source corpus.

Metagenomics projects collect DNA from uncharacterized environments that may contain thousands of species per sample. One main challenge facing metagenomic analysis is phylogenetic classification of raw sequence reads into groups representing the same or similar taxa, a prerequisite for genome assembly and for analyzing the biological diversity of a sample. New sequencing technologies have made metagenomics easier, by making sequencing faster, and more difficult, by producing shorter reads than previous technologies. Classifying sequences from reads as short as 100 base pairs has until now been relatively inaccurate, requiring researchers to use older, long-read technologies. We present Phymm, a classifier for metagenomic data, that has been trained on 539 complete, curated genomes and can accurately classify reads as short as 100 base pairs, a substantial improvement over previous composition-based classification methods. We also describe how combining Phymm with sequence alignment algorithms improves accuracy.

Subfields: Ecology, Molecular Biology

View Paper: https://www.semanticscholar.org/paper/e62f2ee2c117d7df2bc7c74f71e8108b936e475f

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Ecology

In Molecular Biology

Citations in the corpus

Machine learning for metagenomics: methods and tools (2015, methodology): “Phymm (Brady and Salzberg, 2009) uses Interpolated Markov Models to learn variable-length oligonucleotides typical of a given phylogenetic grouping.”
Characterization of the Gut Microbiome Using 16S or Shotgun Metagenomics (2016, methodology): “For shotgun libraries, we compared BLAST top hits, the MEtaGenome ANalyzer MEGAN5, and Metagenomic Phylogenetic Analysis (MetaPhlAn) approaches; however, we do acknowledge that many other excellent tools have also been developed, including PhymmBL (Brady and Salzberg, 2009, 2011), PhyloSift (Darling”
CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes (2015, background): “…complete with ≤5% contamination (Supplemental Table S16), making them excellent reference genomes for analyses such as assigning taxonomy to anonymous genomic fragments (Brady and Salzberg 2009; Parks et al. 2011) or characterizingmetagenomic samples using marker genes (Darling et al. 2014).”
WGSQuikr: Fast Whole-Genome Shotgun Metagenomic Classification (2014, methodology): “Current methods, while relatively accurate, can still take from 8 hours (MetaPhyler [2]) to 4 days (PhymmBL [3]) to analyze a relatively small dataset of 70 thousand 300 bp reads [2].”
pplacer: linear time maximum-likelihood and Bayesian phylogenetic placement of sequences onto a fixed reference tree (2010, methodology): “Phylogenetic placement software sits in a category distinct from taxonomic identiﬁcation software such as MEGAN [10] or Phymm [13].”
Interactive metagenomic visualization in a Web browser (2011, methodology): “When importing classifications from RDP and PhymmBL a color gradient can be used to represent the average reported confidence of assignments to each node.”
Kraken: ultrafast metagenomic sequence classification using exact alignments (2014, methodology): “Table S2: PhymmBL classification accuracy across different confidence levels Confidence Level Accuracy Precision Sensitivity 0.0 54.9 54.9 54.9 (results equal for all confidence levels from 0-0.45) 0.45 54.9 54.9 54.9 0.50 55.2 55.5 54.9 0.55 66.8 86.1 54.5 0.60 69.6 97.5 54.2 0.65 69.7 98.0 54.1 0.”
Centrifuge: rapid and sensitive classification of metagenomic sequences (2016, methodology): “For example, machine-learning-based approaches such as the Naive Bayes Classifier (NBC) (Rosen et al. 2008) and PhymmBL (Brady and Salzberg 2009, 2011) classify <100 reads per minute, which is too slow for data sets that contain millions of reads.”
Metagenomics - a guide from sampling to data analysis (2012, unknown): “There is also number of binning algorithms that consider both composition and similarity, including the programs PhymmBL [55] and MetaCluster [56].”
Performance, Accuracy, and Web Server for Evolutionary Placement of Short Sequence Reads under Maximum Likelihood (2011, unknown): “Recently, Brady and Salzberg (2009) proposed the Phymm and PhymmBL algorithms for metagenomic phylogenetic classification and report an improved classification accuracy for simulated QS compared with BLAST for PhymmBL.”
A deep learning approach to pattern recognition for short DNA sequences (2018, background): “8,9,31,36 11 Despite the recent increase in successful applications of deep learning to genomics, 46-48 to our knowledge no previous work demonstrates that DNNs can classify genomic sequencing data of realistic scales at fine taxonomic resolution, or provides an in-depth comparison to existing…”
VirFinder: a novel k-mer based tool for identifying viral sequences from assembled metagenomic data (2017, methodology): “For example, Glimmer uses interpolated Markov models learned from distributions of k -mers to identify bacterial genes from mixed samples containing eukaryotic sequences [23], and Phymm taxo-nomically classifies bacterial sequences as short as 100 bp using the same type of models [24].”
Chapter 12: Human Microbiome Analysis (2012, background): “Conversely, de novo methods have been developed to directly bin [67,68,69] and call genes within [61,62] metagenomic sequences using DNA features alone (GC content, codon usage, etc.”