SimHyb: a simulation software for the study of the evolution of hybridizing populations. Application to Quercus ilex and Q. suber suggests hybridization could be underestimated

Abstract

We present SIMHYB, a Java-based software for the simulation of mixed hybridizing populations. The software incorporates user-defined initial parameters and input files to account for the initial census size of two species in a closed population, the number of intermediate specific classes, the directional fertility among specific classes, the fitness coefficients for each specific class, the inheritance of fitness, and the degree of ageing and self-incompatibility of the individuals. All these demographic and adaptive parameters can be modified by the user to analyze their effect on the evolution of the mixed population. SIMHYB allows the traceability of each individual, whose pedigree is also recorded. For each simulated generation the software yields an output file that is easily convertible to an input for STRUCTURE, one of the most popular softwares for the Bayesian analysis of populations. Application of SIMHYB to simulate Quercus ilex and Q. suber hybridizing populations, and further analysis with STRUCTURE, reveals that advanced introgressed individuals are very often misclassified with the currently available set of nuclear microsatellite markers, so that introgression between these two species could have been underestimated in previous studies. However, we provide a simple parameter based on STRUCTURE results to identify the directionality of pollination in the progeny of a known mother tree.

Natural hybridisation amongQuercus glabrescens,Q. rugosaandQ. obtusata(Fagaceae): Microsatellites and secondary metabolites markers

“…; Soto et al . ). For this purpose, we evaluate the capacity of the microsatellite set to differentiate between pure and hybrid virtual genotypes created with HybridLab.…”

Molecular evidence of bidirectional introgression between Quercus suber and Quercus ilex

“…A value of K=2 was identified as the optimal K value following the methodology described by Evanno et al (2005) and was used in all subsequent analyses. In order to improve q estimations (Burgarella et al 2009), we included in the STRUCTURE analyses the genotypes of approximately 95 Q. ilex and 96 Q. suber parental trees from the same stand, previously analyzed by Soto et al (2007), along with 59,671 virtual individuals simulated with the software SimHyb (Soto et al 2018). These simulated individuals included a majority of parental trees with some F1 hybrids and backcrosses (Tab.…”

ddRAD Sequencing-Based Identification of Genomic Boundaries and Permeability in Quercus ilex and Q. suber Hybrids

“…Firstly, we checked the performance and accuracy of both classification methods using virtual individuals with known introgression levels. These individuals were simulated with SIMHYB (Soto et al, 2018), based on the allele frequencies of the adult Q. ilex and Q. suber populations. Later on, we estimated contribution of parental species to the genome of each individual in two scenarios of prevalence of hybrids: 1 and 10%.…”

“…The Bayesian approach implemented in STRUCTURE v. 2.3.4 (Pritchard et al, 2000), as well as the hybrid index provided by the INTROGRESS R-package (Gompert and Buerkle, 2010), were then used to classify individuals according to their introgression levels. Firstly, and following the procedure established by Burgarella et al (2009), accuracy of the classification was evaluated using virtual individuals of known pedigree created with SIMHYB (Soto et al, 2018). ScnIV provided a fairly accurate classification of virtual hybrid individuals.…”

“…In 2009, current hybridization was estimated in less than 2%, in a study using eight nSSR and including natural pure and mixed stands, as well as provenance trials covering the whole distribution range of Q. suber ( Burgarella et al., 2009 ). Nevertheless, recent studies have shown that classification/identification power of the set of nSSR used is too low, and thus, hybridization and introgression could have been underestimated ( López de Heredia et al., 2018a ; Soto et al., 2018 ). Therefore, more powerful marker sets are needed in order to obtain accurate estimations.…”