Related Experiment Videos
Towards building the tree of life: a simulation study for all angiosperm genera
Nicolas Salamin1, Trevor R Hodkinson, Vincent Savolainen Coates
1Department of Botany, University of Dublin, Trinity College, Dublin 2, Ireland. nicolas.salamin@unil.ch
Systematic Biology
|July 14, 2005
Summary
Building large phylogenetic trees for evolutionary studies is feasible. Tree reconstruction accuracy plateaus with ~1000 nucleotides, and more taxa don't significantly reduce accuracy, with branch length distribution being key.
Area of Science:
- Phylogenetics
- Computational Biology
- Evolutionary Biology
Background:
- Comprehensive phylogenetic trees are crucial for understanding evolutionary processes.
- Reconstructing large-scale evolutionary trees, especially for the Tree of Life, involves sampling millions of taxa.
Purpose of the Study:
- To assess the performance of phylogenetic reconstruction methods for large datasets.
- To determine factors influencing the accuracy of inferring large phylogenetic trees.
Main Methods:
- Monte Carlo simulations using parameters from large angiosperm DNA matrices (141-567 taxa).
- Simulated DNA matrices with varying sequence lengths (100-10,000 bp) under the HKY85+G model.
- Analyzed simulated data using Maximum Parsimony and Neighbor Joining methods.
Main Results:
- Phylogenetic reconstruction accuracy rapidly increased with sequence length, plateauing around 1000 nucleotides.
- Increasing the number of taxa (141 to 567) did not significantly decrease accuracy.
- Branch length distribution was more critical for accurate large tree inference than the rate of evolution.
Conclusions:
- Phylogenetic reconstruction methods are effective even for very large trees (simulated 13,000 taxa).
- Heuristic searches can accurately infer a high percentage of nodes in massive phylogenetic trees.
- The study provides insights into the scalability and key factors for reconstructing large evolutionary trees.