Visual Analytics for the Exploration of Phylogenetic and Topographic-Units Supertrees
In biological sciences, dendrograms - or trees - are useful to present information on evolutionary relationships and biogeographical history in a hierarchical format. Defined as acyclic rooted or unrooted graphs, trees may be based on genetic, molecular, morphological, geological, behavioral or geographical pieces of evidence. The trees estimation process is often computationally intensive. In large scope studies, when one is dealing with many different biological groups or complex biogeographical history, techniques such as supertree methods may be used to create a large tree that combines convergent nodes of a set of small trees. Despite the many developed methods for supertree construction, Matrix Representation with Parsimony (MRP) is still extensively used to transform each single separated tree into a binary absence/presence matrix. However, the MRP procedure has limitations: 1) lack of an easy-to-use and easy-to-access automated tool to prepare matrix representations from the original topologies, as well as the combined matrix; and 2) the lack of truly interactive visualization techniques to explore the resultant supertree and to apply further analysis without the need to start the process from the beginning. In this work, we proposed a solution for the two mentioned limitations that relies on the concept of Visual Analytics. Motivated by the current procedure of supertree construction and discussions with domain experts, we present a preliminary requirements study, a general framework for two visual analytics approaches, and the developed parts of both approaches. The first is the interactive phylogenetic supertree (iPhyloS), which deals with large scope presentations of evolutionary relationships among biological groups, while the second is the interactive topographic-units supertree (iTUS), which deals with species endemism data. We developed and tested the data analysis components of iPhyloS and iTUS in the form of two stand-alone online tools: 1) Building combined MRP-matrices (BuM), for iPhyloS, and 2) interactive Topographic-Unit Parsimony Analysis (iTUPA), for iTUS. BuM is an online tool for dealing with the preparation of the MRP-matrices of multiple phylogenetic trees and combining the resulting data into a single supermatrix. BuM has shown itself a reliable tool for building combined MRP-matrices from original topologies. iTUPA is an online automated application for generating species matrices of occurrences versus previously defined topographic units and analyzing them through a parsimony approach. iTUPA allows interactively visualize the endemism data and the previously defined topographic with the Google Maps API. Both iPhyloS and iTUS has an interactive visualization component which is different for each system. We have integrated phylotree.js into our developed tools to create a prototype of the supertree system. This library will be further enhanced to fulfill all of the defined requirements, and to reach the final goal of developing a whole system, based on visual analytics, for exploring phylogenetic and topographic-units supertrees.