What does Mesquite do?
(updated March 2001)
Mesquite was built for comparative studies of organisms, genes, and their evolution, focusing on:
- Taxa - These are the sampled entities being compared, typically either species or genes. Thus, if a taxon is a species, then its characters are the features we observe among its individuals, and the trees we use to represent the species relationships are phylogenetic trees. If the taxon is a gene, then its characters are (typically) its nucleotide positions or amino acid residues, and the trees representing the gene relationships are gene trees. (In Mesquite, "taxon" and "taxa" refer to terminal taxon and taxa.)
- Characters - These are the observed or measured characteristics or variables of the taxa. These could represent nucleotide sequences, morphometric measurements, ecological parameters of the species, and so on.
- Trees - Trees that represent the relationships among the taxa could be phylogenetic trees (if the taxa are species) or gene trees (if the taxa are gene sequences).
Mesquite's analyses often compare taxa via their characters, using the relationships among taxa (trees) as a guide, in order to make inferences about evolutionary history or process. It can also study taxa and their trees (without reference to characters) and taxa and their characters (with reference to trees).
What questions does Mesquite answer? What functions does it perform?
Because the Mesquite system is extensible, we cannot predict all of the biological questions Mesquite might answer. Below is a list of the topics for which Mesquite could supply analyses. At least some work has been done to develop modules in each of these areas, although these capabilities might not be represented in any particular installation of Mesquite. Also, some of these areas have received little effort, and thus there are plenty of opportunities for other programmers to contribute modules. This list is more a map of where we can see Mesquite developing than of what has already been done.
Links are provided to some packages of modules that were in development at the time of this writing, although these links will not work if you don't have the packages installed in your Mesquite system.
- Interpreting characters evolving on trees: What is the history of character evolution? How often have characters changed, at what rates, with what biases, and with what correlations?
- Parsimony interpretations of character evolution (categorical, molecular and continuous data) [see Parsimony package]
- Reconstructions of ancestral states
- Likelihood interpretations of character evolution [prototypes, but no package available]
- Likelihood of tree under various models of molecular evolution
- Estimation of parameters of change
- Reconstructions of ancestral states
- Correlated character evolution
- Choosing trees: What tree is best supported by data, or how do trees differ in their support by the data?
- Interaction with tree search programs like PAUP* [see PAUP* communication package]
- Interactive manipulations of trees with feedback on parsimony, likelihood and other measures of the tree [see Basic package]
- Distance-based analyses
- Tree searching or tree filtering using any available measure for trees, including fit of gene tree into species tree, measured character correlations, and so on. [see Tree search package]
- Comparing trees and tree shape
- Simulations and hypothesis testing: With a particular stochastic process of speciation and extinction, or character evolution, or gene coalescence (perhaps constrained by a hypothetical species phylogeny or population history), what outcomes do we see according to simulations?
- Simulated speciation (birth/death process trees)
- Coalescence simulations of genes within populations [see Coalescence package]
- Molecular evolution
- Likelihood calculations for sequence evolution
- Simulations of sequence evolution
- Sequence statistics (e.g., GC vs. AT bias)
- Population genetics and coalescence [see Coalescence package]
- Simulation of gene copy coalescence within a population
- Coalescence simulation within a species tree
- Fit of gene tree into species tree (statistics, visualization)
- Multivariate analysis and morphometrics [see Rhetenor package]
- Plotting a phylogenetic tree into character space (2D, 3D with rotation)
- Principal Components, Canonical Variates Analyses
- Geometric morphometrics
- Landmark data
- Reconstruction of ancestral forms on a tree
- thin plate spline deformations
- Interactions with other programming efforts
Interface, Facilities and Utilities
Mesquite is an interactive system using a graphical user interface, although it can be scripted. It presents windows, menus, buttons, and tools for manipulating data and performing analyses. Some of its features to manage data and assumptions are outlined here.
- Files and file formats
- Data files can contain more than one set of taxa, more than one data matrix, more than one trees block
- NEXUS format used as standard format
- Import and/or export facilities for PHYLIP, Hennig86-NONA, Clustal, NBRF files
- Character data and assumptions
- Spreadsheet editor for categorical, molecular and continuous data
- List windows to assign models of evolution, weights, exclusion, and other properties to characters
- Histograms and scattergrams showing values calculated over multiple characters or character matrices
- Trees manipulation and management
- Tree window in which trees can be viewed, analyzed, manipulated, saved
- Specialized tree windows to compare multiple trees or a single tree with different analyses compared
- Histograms and scattergrams showing values calculated over multiple trees or blocks of trees
- Scripting language allows most user commands to be scripted
- Script written automatically to file to return analysis to current state
- Help and auto-documentation
- Explanation area in most windows to explain commands and the current state of objects
- Automatic generation of documentation summarizing the currently available features
© W. Maddison & D. Maddison 2001