TreeCluster: Clustering biological sequences using phylogenetic trees

Abstract

Clustering homologous sequences based on their similarity is a problem that appears in many bioinformatics applications. The fact that sequences cluster is ultimately the result of their phylogenetic relationships. Despite this observation and the natural ways in which a tree can define clusters, most applications of sequence clustering do not use a phylogenetic tree and instead operate on pairwise sequence distances. Due to advances in large-scale phylogenetic inference, we argue that tree-based clustering is under-utilized. We define a family of optimization problems that, given an arbitrary tree, return the minimum number of clusters such that all clusters adhere to constraints on their heterogeneity. We study three specific constraints, limiting (1) the diameter of each cluster, (2) the sum of its branch lengths, or (3) chains of pairwise distances. These three problems can be solved in time that increases linearly with the size of the tree, and for two of the three criteria, the algorithms have been known in the theoretical computer scientist literature. We implement these algorithms in a tool called TreeCluster, which we test on three applications: OTU clustering for microbiome data, HIV transmission clustering, and divide-and-conquer multiple sequence alignment. We show that, by using tree-based distances, TreeCluster generates more internally consistent clusters than alternatives and improves the effectiveness of downstream applications. TreeCluster is available at https://github.com/niemasd/TreeCluster .

Global Distribution and Evolution of Mycobacterium bovis Lineages

“…Phylogenetic clustering was performed using TreeCluster software (Balaban et al, 2019), which is a phylogeny-based clustering method. The ML phylogenetic tree of the reduced M. bovis dataset (n = 1,201) was used as input and the method "Avg Clade" was applied, which means that the average pairwise distance between leaves in a cluster should be at most "t", where we used t = 0.012 substitutions/site.…”

“…Deletions and SNP correspond to the clonal complex markers European 1 (Eu1), European 2 (Eu2) and African 1 (Af1), indicated by triangles. Seven M. bovis clusters (2 through 8) were identified using TreeCluster (Balaban et al, 2019) and they are indicated by colored tips. Cluster 1 corresponds to the outgroup composed of M. orygis and M. caprae genomes.…”

“…It has been recently shown that phylogeny-based clustering of populations can better reflect divergence and similarity between microorganisms when compared to more simplistic pairwise distance-based methods (Ragonnet-Cronin et al, 2013;Balaban et al, 2019). By taking into account phylogenetic trees, phylogeny-based clustering methods leverage from model-based, statistically rigorous corrections of sequence distances depicted in branch lengths (Balaban et al, 2019). Thus, we used TreeCluster, a phylogeny-based clustering method, to better understand the populational structure of M. bovis genomes.…”

Review of Hepatitis E Virus in Rats: Evident Risk of Species Orthohepevirus C to Human Zoonotic Infection and Disease

“…With a comparison, the mean p-distance scores are 9.7%, 28.9%, 19.7%, 17.4%, 23%, 15.1%, 9.7% within HEV-1, HEV-2, HEV-3, HEV-4, HEV-6, HEV-7, and HEV-8 groups in Orthohepevirus A , respectively. The cluster analysis is a common technique to classify or grouping data (sequences) to provide evolutionary connections or map the relationship between the related species [ 69 , 70 ]. The rodent hepeviruses form a discrete phylogenetic cluster among species Orthohepevirus C in family Hepeviridae ( Figure 3 ).…”

In silico comparative genomics of SARS-CoV-2 to determine the source and diversity of the pathogen in Bangladesh

“…Sequences were then put into different clades based on specific mutations proposed in GISAID (23) and further classified as D614G type (24,25). Subsequently, another phylogenetic tree and haplotype network containing only SARS-CoV-2 sequences from Bangladeshi was constructed and categorized using the same tools, and additionally one step further clustered with TreeCluster (26). The direction of selection in sequences from Bangladesh was calculated by the SLAC algorithm (27) in the Datamonkey server (28).…”