Although Burgess and Yang (2008) modeled sequencing errors and violations of the molecular clock, those factors were found to have only minor impact on estimation of parameters concerning the human, chimpanzee, and gorilla in the analysis of the curated data (compare tables 2 and 5 in Burgess and Yang 2008).
With respect to the molecular clock, there are many factors leading to uncertainty.
The Bayesian skyline and relaxed clock models were selected over demographic (constant and exponential growth) and strict molecular clock models using Bayes factor analysis [30] (support >20 at log10 Bayes factor scale).
Because one of the purposes of this study was to infer historical relationships with respect to divergence times, a molecular clock was tested using Bayes factors.
We carried out Bayesian phylogenetic analysis of the entire dataset in BEAST v. 1.7.4 [ 7– 9], using the GTR+ Γ nucleotide substitution model and uncorrelated lognormal relaxed molecular clock, which had greater Bayes factor support than a strict clock (BF = 3.65) [ 10, 11].
As with 8-taxon simulated datasets, we tested whether the molecular clock could be rejected using Bayes factors and the likelihood ratio test.
The question appears whether or not proinflammatory factors can influence molecular clock mRNA.
Various different substitution, coalescent and molecular clock models were compared by calculating Bayes Factors (BF), which is the difference in log marginal likelihoods between two model combinations [ 25, 26].
Bayes factor tests of a molecular clock analysis indicate that a relaxed lognormal clock fits the data significantly better than a strict clock model.
Non-histone substrates of SIRT1, as found in C2C12 myotubes, include regulatory molecules that modulate energy metabolism, such as PPARγ and PGC-1α [ 114], key factors that regulate the core molecular clock.
