This method is much faster than the direct optimization methods under high evolution rearrangement event rate, but the accuracy of the reconstructed gene order is lower than that of the direct optimization methods.
We then report the results of a simulation study, in which we test our method under various evolution scenarios, and compare it with the composite two-marker method in [ 18] and the multimarker methods in [ 21, 26, 27].
The EM algorithm was applied to the set of 391 orthologous genes from 19 eukaryotic species ([ 20] and see Methods), under the homogeneous evolution assumption, and the profile likelihood technique was used to estimate the 95% confidence interval for each parameter (Additional file 1).
The suggestions of positive selection of the several site-specific analyses are summarized in table 2, where only sites with more than one method suggesting evolution under positive selection are shown (complete results are in supplementary table S3, Supplementary Material online).
The second way is to mimic another powerful method of evolution -- sexual reproduction.
In Fig. 2 we report a plot of the probability P dN/dS>1) that a codon has dN/dS>1 for rrgA, and sites for which the Bayes-Empirical-Bayes inference (BEB) method [26] (see Materials and Methods section) supports evolution under positive selection are evidenced.
However, inference biases are considerable for both methods under particular departures from stationarity (i.e., when adaptive evolution is prevalent).
Two methods under consideration are a).
An e-value cutoff of e−10 was selected to strike a balance between statistical significance and the detection of remote homology ('Materials and methods' under 'Evolution analysis').
We identified a set of twenty signature genes that were highly expressed in each of the 26 hematopoietic populations (Benita et al., 2010; Seita et al., 2012; 'Materials and methods' under 'Evolution analysis').
All B. schlosseri candidate protein-coding genes were compared to human and mouse proteomes (UniProtKB/Swiss-Prot; see Sequence data in 'Materials and methods ' under 'Phylogenomic analyses') using blastp and an e-value threshold of e−10 (see Sequence Data, 'Materials and methods' under 'Evolution analysis').
