GLOBAL OVERVIEW
HERMES version: 3.0
The HERMES index is a method of quantifying molecular evolution of mitochondrial genomes in different species and clusters; this method was originally proposed in Plazzi et al. (2016). The index relies on maximum likelihood factor analysis to summarize different measures that are typically found to be linked with evolutionary rates; it is intended to be computed a posteriori, i.e. after a phylogenetic and genomic analysis. As different empirical measures are merged together in a single score, it is a “hyper-empirical” index; moreover, it is a relative measure, because all species are compared with an outgroup: therefore, it was called Hyper-Empirical Relative Mitochondrial Evolutionary Speed (HERMES) index.
The present Python script performs data collection and then calls a dedicated R script to complete the factor analysis. The mitogenomic features that are currently implemented in HERMES-v3.0.py are:
the percentage of Unassigned Regions (URs);
the Amount of Mitochondrial Identical Gene Arrangements (AMIGA) index;
the absolute value of the Strand Usage (SU) skew;
the root-to-tip distance computed over a given phylogenetic tree;
the Maximum Likelihood (ML) pairwise distance from a given outgroup;
the AT content;
the AT skew;
the GC skew;
the number of (annotated) genes;
the length of the molecule;
the Codon Adaptation Index (CAI), as defined in Sharp and Li (1987) and Xia (2007);
the topology (either circular or linear);
the UR AT content;
the UR median length.
For each possible combination of at least two of these variables, a factor analysis is carried out. Normalization and varimax rotation are used, factor scores are found using correlation preserving, and correlations are found using the Pearson method; given the possible presence of missing values, missing data are set to be imputed using the median.
All the variables are pooled together for each species into the value of a single loading: we define this score as the HERMES score of a given species.
The best-performing variable set and the goodness-of-fit of the analysis is assessed following the recommendations of Hu and Bentler (1999): Tucker-Lewis Index (TLI) greater than 0.95; root mean square of the residuals (SRMR) smaller than 0.08; root mean squared error of approximation (RMSEA) less than 0.06; moreover, the Kaiser-Meyer-Olkin index (KMO) is taken into account on this regard.
REFERENCES
Hu L-T, Bentler PM. 1999. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal. 6:1-55.
Plazzi F, Puccio G, Passamonti M. 2016. Comparative Large-Scale Mitogenomics Evidences Clade-Specific Evolutionary Trends in Mitochondrial DNAs of Bivalvia. Genome Biol Evol. 8:2544-2564.
Sharp PM, Li WH. 1987. The Codon Adaptation Index--a Measure of Directional Synonymous Codon Usage Bias, and Its Potential Applications. Nucleic Acids Res. 15:1281-1295.
Xia X. 2007. An Improved Implementation of Codon Adaptation Index. Evol Bioinform. 3:53-58.