• Français
  • English

MDA

MDA: a MATLAB-based program for morphospace-disparity analysis [ Science Direct ]

 

Developed by Nicolas Navarro

 

The great development of morphospace-disparity analysis in paleobiological studies, and the challenge of extending them to other disciplines, require the development and use of various complementary metrics. It would allow to apprehend state-space patterns in terms of features such as dispersion, aggregation and location in order to extract complementary quantitative information about the way state-space is structured.
MDA is a MATLAB® program that examines patterns of state-space occupation. Four subroutines are available with which to visualize morphospace patterns (i) in terms of their features such as dispersion, aggregation and location, thereby allowing users to extract complementary quantitative information about how the state-space is structured, and (ii) in terms of changes in those patterns that can be compared with other biotic (e.g., extinction, origination rates) or abiotic (e.g., environmental proxy) information. The program incorporates many of the latest and most widely used statistical parameters for describing multivariate spaces. The parameters are estimated on the basis of bootstrap resampling or bootstrap rarefaction procedures. The help folder contains input data files and results based on stochastic simulation of the evolution of monophyletic clade (using m-file also contained in this folder) so as to illustrate the program’s various options. The versatility of MDA means that the most interesting patterns can be extracted rapidly from data and that the program can be applied readily to a wide range of state-space problems.

 

MDA concatenates some m-files developed in the aim to realize disparity analysis. This concatenation in one main function has been imposed for simple utilization by unfamiliar users of MATLAB. Input data are loaded using either text command or dialog box. Some shortcuts were introduced using default filenames (in this case, data files will be stored in the program folder). No additional toolbox are required. This program is copyrighted by N. Navarro but you can freely used, distributed or modified it (see copyright information in the preamble of the program). This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRENTLY.
Please cite when use under the form:
Navarro, N. 2001. MDA: Morphospace-Disparity Analysis for Matlab. Version 1.2. Biogéosciences-Dijon Publisher
Or the introductory paper:
Navarro, N. in press. MDA: a MATLAB-based program for Morphospace-Disparity Analysis. Computers & Geosciences.

 

 

>>> Download MDAvs1.2.03.zip

 

 

Metrics and Analysis performed by MDA

 

10 estimates of disparity are performed:

– 5 range-based: Sum of Ranges, Root Product of Ranges, Max Distance, PCO-Volume, Area of the Convex Hull (first 2-D);

– 5 variance-based: Sum of Variances, Root Product of Variances, Mean Pairwise Distance, Median Pairwise Distance, Mean Distance to group centroid;

 

 

2 estimates of dispersion:

– Minimum and Maximum (but limited to first 10 dimensions).

 

 

4 types of analysis are performed:

– disparity analysis of single group (SGA),

– multiple group (MGA),

– partial disparity analysis (PDA),

– bootstrap tail test (BTailTest).

 

Another function (MDA_1D) analyzed one dimensional data (a max of 10 dimensions are simultaneously analyzed): estimation of univariate estimates of dispersion (variance and range) and estimates of location (mean, mininum and maximum).Estimates are obtained by bootstrap or rarefaction using bootstrap (if metric is biased by sample size).  Many options are available using an user text-interface: size of rarefaction, number of bootstrap replications, level of confidence interval.

 

See « User Guide to MDA » (this pdf file is contained in the zip archive of MDA).

See the introductory article of MDA submitted to Computers & Geosciences

 

 

Analysis example: the user-interface and the graphical outputs

Analysis preamble

1st step: input dialog box

text user-interface

Disparity metrics (SGA-MGA)

Location estimates Min & Max on each PC

Bootstrap distribution (BTailTest)

 

 

Disparity: some references

Here some works about disparity: EndNote format; BibTex format; txt (it is not an exhaustive list)

Log In

Create an account