Multidisciplinary Challenges in the Educational Assessment and Identification of Speech and Language Disorders in Türkiye: Perspectives from Guidance and Research Centers.

Mutation and recombination are important evolutionary processes that shape levels and patterns of genetic diversity in populations. Germline mutations are the ultimate source of novel genetic variation, whilst recombination shuffles this variation into potentially novel haplotypes via crossover and non-crossover events. The rate of input of new mutations, as well as the rate of recombination events, have been shown to vary at every level of measurement: across the Tree of Life, between and within species, and across the genome (for mutation rate variation, see the reviews of Baer et al. , Lynch , Hodgkinson and Eyre-Walker , Pfeifer , for recombination rate variation, see the reviews of Ritz et al. , Stapley et al. , Johnston ). Both mutation and recombination rate estimation can be performed either via direct observation from pedigrees or indirectly from sequenced population samples (though classical disease-incidence approaches have also historically been utilized for mutation rate estimation in humans; Haldane , ). The direct estimation of both processes relies on high-throughput genome sequencing of parent-offspring trios or multi-generation pedigrees, counting the number of de novo mutations as well as crossover and non-crossover events that have occurred from one generation to the next (see the review of Pfeifer for an overview, and Pfeifer , Bergeron et al. for a discussion of the challenges in direct rate estimations). Due to the rarity of both spontaneous mutations and meiotic exchange events in vertebrates, resolution with such direct estimation approaches is relatively coarse, given the small number of generations generally considered (see the review of Clark et al. ). Consequently, they provide a genome-wide rate estimate of mutation and recombination, as opposed to a fine-scale map of rate heterogeneities across the genome that is necessary for a variety of applications, including genome-wide association studies and selection scans. By contrast, indirect mutation and recombination rate estimation are performed on species-level divergence data and population-level polymorphism data, respectively. Central to such indirect mutation rate estimation approaches is the observation that the neutral mutation rate is equal to the neutral divergence rate (Kimura , ), with the number of substitutions that accumulate in a lineage being proportional to the per-generation mutation rate. Thus, historically-averaged mutation rates across the divergence time between the target species and an outgroup species can be inferred from phylogenetic sequence data in neutral genomic windows, thereby generating a fine-scale genomic map of mutation rate heterogeneity. However, there is often great uncertainty in both the generation time of a species and the divergence times between the species under investigation. Estimated mutation rates are therefore given across a range of likely generation and divergence times in order to span this uncertainty. Instead of divergence data, indirect recombination rate estimation approaches rely on population-level data of unrelated individuals for the inference of historical recombination rates from observed patterns of linkage disequilibrium (LD; see the reviews of Stumpf and McVean , Peñalba and Wolf ), again utilizing neutral genomic windows to generate fine-scale maps across the genome. These inferred rates are necessarily sex-averaged, and one must account for other population genetic processes that can alter LD (e.g., selection and population history; Dapper and Payseur , Samuk and Noor ) and thus potentially confound recombination rate inference. To limit the impact of such confounding factors on the indirect inference of both mutation and recombination rates, high-quality genome annotations are necessary to identify regions of the genome that are evolving neutrally; additionally, a well-fitting demographic model is necessary in the case of recombination rate inference (Johri et al. , ).