GHOST: Recovering Historical Signal from Heterotachously Evolved Sequence Alignments

Crotty, Stephen; Minh, Bùi Quang; Bean, Nigel; Holland, Barbara R.; Tuke, Jonathan; Jermiin, Lars S.; Haeseler, Arndt von

doi:10.1093/sysbio/syz051

Cited by 116 publications

(130 citation statements)

References 56 publications

Supporting

Mentioning

127

Contrasting

Order By: Relevance

“…For efficiency, we restricted the analyses to a subset of models (“‐mset LG” for amino acids, “‐mset GTR” for nucleotides) and employed the relaxed hierarchical clustering algorithm “‐rcluster 10” (Lanfear, Calcott, Kainer, Mayer, & Stamatakis, 2014). Heterotachy ML analyses used the GHOST (General Heterogeneous evolution On a Single Topology, Crotty et al, 2019) model to address rate variation across sites and lineages (Lopez, Casane, & Philippe, 2002). Node support values were calculated using 1,000 sh‐alrt replicates (Guindon et al, 2010) and 1,000 ultrafast bootstraps (Hoang, Chernomor, von Haeseler, Minh, & Vinh, 2018).…”

Section: Methodsmentioning

confidence: 99%

Streamlining universal single‐copy orthologue and ultraconserved element design: A case study in Collembola

Sun

Ding

Orr

et al. 2020

Molecular Ecology Resources

View full text Add to dashboard Cite

Genomic data sets are increasingly central to ecological and evolutionary biology, but far fewer resources are available for invertebrates. Powerful new computational tools and the rapidly decreasing cost of Illumina sequencing are beginning to change this, enabling rapid genome assembly and reference marker extraction. We have developed and tested a practical workflow for developing genomic resources in nonmodel groups with real‐world data on Collembola (springtails), one of the most dominant soil animals on Earth. We designed universal molecular marker sets, single‐copy orthologues (BUSCOs) and ultraconserved elements (UCEs), using three existing and 11 newly generated genomes. Both marker types were tested in silico via marker capture success and phylogenetic performance. The new genomes were assembled with Illumina short reads and 9,585‒14,743 protein‐coding genes were predicted with ab initio and protein homology evidence. We identified 1,997 benchmarking universal single‐copy orthologues (BUSCOs) across 14 genomes and created and assessed a custom BUSCO data set for extracting single‐copy genes. We also developed a new UCE probe set containing 46,087 baits targeting 1,885 loci. We successfully captured 1,437‒1,865 BUSCOs and 975‒1,186 UCEs across 14 genomes. Phylogenomic reconstructions using these markers proved robust, giving new insight on deep‐time collembolan relationships. Our study demonstrates the feasibility of generating thousands of universal markers from highly efficient whole‐genome sequencing, providing a valuable resource for genome‐scale investigations in evolutionary biology and ecology.

show abstract

Section: Methodsmentioning

confidence: 99%

Streamlining universal single‐copy orthologue and ultraconserved element design: A case study in Collembola

Sun

Ding

Orr

et al. 2020

Molecular Ecology Resources

View full text Add to dashboard Cite

show abstract

“…IQ-TREE 2 offers a number of advanced models for phylogenomic data including partitioned models (Lanfear et al 2012;Chernomor et al 2016), mixture models (Le et al 2008;Le and Gascuel 2010;Le et al 2012), posterior-mean site frequency models (Wang et al 2017), and heterotachy models (Crotty et al 2019). For allele frequency data IQ-TREE 2 implements the polymorphism-aware models (Schrempf et al 2016;Schrempf et al 2019).…”

Section: Time-reversible Models Of Sequence Evolutionmentioning

confidence: 99%

IQ-TREE 2: New models and efficient methods for phylogenetic inference in the genomic era

Minh

Schmidt

Chernomor

et al. 2019

Preprint

Self Cite

1,181

1,435

View full text Add to dashboard Cite

IQ-TREE (http://www.iqtree.org) is a user-friendly and widely used software package for phylogenetic inference using maximum likelihood. Since the release of version 1 in 2014, we have continuously expanded IQ-TREE to integrate a plethora of new models of sequence evolution and efficient computational approaches of phylogenetic inference to deal with genomic data. Here, we describe notable features of IQ-TREE version 2 and highlight the key advantages over other software.

show abstract

“…Unlike the trees of the 3-class model, the trees of the 10-class model (Supplementary Figure S3) preferred by AIC bear the hallmarks of overfitting, as described in Crotty et al (2017). Many of the classes bear strong similarity to each other, most noticeably the first three classes all appear strongly conserved across all taxa, much like the first class of the 3-class model.…”

Section: Resultsmentioning

confidence: 99%

“…We used IQ-TREE (Nguyen et al, 2015) to fit a GHOST model to the sequence alignment. We used the model selection procedure outlined in Crotty et al (2017) to choose the model of sequence evolution and number of classes. After performing the inference we analysed the site-wise probabilities of evolving under each inferred class to identify influential sites, contiguous regions and genes within the genome.…”

Section: Methodsmentioning

confidence: 99%

“…We made use of two new bioinformatics tools to analyse a sequence alignment consisting of 29 whole virus genomes, 14 CBSV and 15 UCBSV. We first performed phylogenetic inference under the GHOST model of sequence evolution (Crotty et al, 2017). The GHOST model is a mixture model which enables the fitting of multiple classes to a single alignment.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Characterising Genetic Diversity in Cassava Brown Streak Virus

Crotty

Rohrlach²,

Ndunguru

et al. 2018

Preprint

Self Cite

View full text Add to dashboard Cite

Plant viruses represent a significant threat to food security for many 23 global populations. Cassava Brown Streak Virus (CBSV) causes immense damage 24 to cassava crops in Eastern, Central and Southern Africa. The eradication of CBSV 25 is a difficult challenge, as it has been shown to be fast-evolving and it is 26 transmitted by flying insects that are ubiquitous in cassava growing regions. In this 27 paper we demonstrate the ability of two new developments in bioinformatics that 28 can be used to increase our understanding of CBSV and ultimately inform 29 strategies for its combat. We reconstruct the phylogeny of 29 whole-genome virus 30 isolates using the GHOST model. This phylogeny identifies three distinct clades 31 among the viruses and highlights a section of the genomes that is highly influential 32 in their divergence. We also perform Multiple Correspondence Analysis on the 33 alignment which is consistent in recovering the three clades, and offers insight on 34 the significance of the influence of a variety of external variables on the evolution of 35 the viruses. Knowledge and information from this analysis will be used as a base on 36 which to formulate sustainable Cassava Brown Streak Disease (CBSD) 37 management strategies in Africa. 38Introduction 39 It is difficult to overemphasize the importance of agriculture to people living in 40 3 sub-Saharan Africa. The population of sub-Saharan Africa exceeds one billion 41 people, and more than 60% of the population live in rural areas (United Nations, 42 2015). Beyond the obvious benefit of nutrition, agriculture is also the primary 43 source of income for the majority of sub-Saharan African families, and so access to 44 essentials such as health care and education depend indirectly on agriculture. As 45 such, the productivity of the agricultural industry in a given country and year is a 46 key indicator of the health and wellbeing of the population and the economy. 47 One of the major threats to agricultural productivity comes in the form of 48 crop pests and diseases. Of particular significance are the viruses of the genus 49 Ipomovirus. They have the potential to devastate crops, and they infect some of 50 the most economically important, commonly grown food staples across sub-Saharan 51 Africa, such as cassava and sweet potato. Cassava Brown Streak Disease (CBSD) is 52 caused by two closely related Ipomovirus species, Cassava Brown Streak Virus 53 (CBSV) and Ugandan Cassava Brown Streak Virus (UCBSV). CBSD is widespread 54 and causes significant reduction in both the quality and the yield of cassava crops, 55 making it a strong barrier to food security and economic prosperity in the regions 56 in which it is grown. 57 Efforts to combat CBSV and UCBSV are ongoing, but there are many 58 challenges. The viruses are transmitted between plants courtesy of whiteflies 59 4 (Bemesia tabaci ) and other potential vectors (Ateka et al., 2017), so it is critical 60 that infected plants are removed from crops swiftly. This indicates a critical need 61 fo...

show abstract

GHOST: Recovering Historical Signal from Heterotachously Evolved Sequence Alignments

Cited by 116 publications

References 56 publications

Streamlining universal single‐copy orthologue and ultraconserved element design: A case study in Collembola

Streamlining universal single‐copy orthologue and ultraconserved element design: A case study in Collembola

IQ-TREE 2: New models and efficient methods for phylogenetic inference in the genomic era

Characterising Genetic Diversity in Cassava Brown Streak Virus

Contact Info

Product

Resources

About