Probabilistic models and their impact on the accuracy of reconstructed ancestral protein sequences

Tal Pupko*, Adi Doron-Faigenboim, David A. Liberles, Gina M. Cannarozzi

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

Abstract

Modeling of sequence evolution is fundamental to ancestral sequence reconstruction. Care must be taken in choosing a model, however, as the use of unrealistic models can lead to erroneous conclusions. The choice of model and the effects of assumptions inherent within are discussed in this chapter in terms of their effects on probabilistic ancestral sequence reconstruction. This chapter discusses standard probabilistic models, site rate variation to these models, and deviations from the standard (homogeneous, stationary, reversible) models. Model selection, selecting one model from many, given data, and the comparison of different models are included as well as covarion models, the use of outside information when modeling, and the treatment of gaps.

Original languageEnglish
Title of host publicationAncestral Sequence Reconstruction
PublisherOxford University Press
ISBN (Electronic)9780191714979
ISBN (Print)9780199299188
DOIs
StatePublished - 1 Sep 2008

Keywords

  • Branch lengths
  • Covarion
  • Gaps
  • Likelihood
  • Models
  • Parsimony
  • Protein structure
  • Rate variation

Fingerprint

Dive into the research topics of 'Probabilistic models and their impact on the accuracy of reconstructed ancestral protein sequences'. Together they form a unique fingerprint.

Cite this