The Role of Epistasis in the Manifestation of Heterosis: A Systems-Oriented Approach

Abstract

Heterosis is widely used in breeding, but the genetic basis of this biological phenomenon has not been elucidated. We postulate that additive and dominance genetic effects as well as two-locus interactions estimated in classical QTL analyses are not sufficient for quantifying the contributions of QTL to heterosis. A general theoretical framework for determining the contributions of different types of genetic effects to heterosis was developed. Additive 3 additive epistatic interactions of individual loci with the entire genetic background were identified as a major component of midparent heterosis. On the basis of these findings we defined a new type of heterotic effect denoted as augmented dominance effect d i * that comprises the dominance effect at each QTL minus half the sum of additive 3 additive interactions with all other QTL. We demonstrate that genotypic expectations of QTL effects obtained from analyses with the design III using testcrosses of recombinant inbred lines and composite-interval mapping precisely equal genotypic expectations of midparent heterosis, thus identifying genomic regions relevant for expression of heterosis. The theory for QTL mapping of multiple traits is extended to the simultaneous mapping of newly defined genetic effects to improve the power of QTL detection and distinguish between dominance and overdominance.

Genome-wide mapping of QTL associated with heterosis in the RIL-based NCIII design

“…is the augmented additive effect of QTL k, which measures the net contribution of the QTL to the parental difference (PD) [14,36] …”

“…is the augmented dominance effect of QTL k, which measures the net contribution of the QTL to the midparent heterosis (MPH) [14,36]; and (Table 1). Models (3) and (6) were working models for our QTL mapping approach in the RIL-based NCIII design under the F ∞ metric model.…”

“…Heterosis represents one of the most revolutionary advancements in crop improvement, i.e., maize [8] and rice [4]. In the genetic dissection of heterosis, NCIII design, proposed by Comstock and Robinson [9,10], is one of the most powerful and widely used mating schemes [11][12][13][14][15][16][17]. Therefore, the importance of heterosis necessitates an in-depth study of the NCIII design.…”

“…The reasons are as follows. First, RIL is a permanent segregating population, so it can be shared between research groups, and testcross progeny can be repeatedly generated and phenotyped as needed [14,15,26,29]. Second, RIL allows more recombinational segregation of tightly linked QTL than the other populations, and this offers a higher mapping resolution [15,22,29].…”

“…Melchinger et al [14] developed a general theoretical framework of composite interval mapping for determining the contributions of different types of genetic effects to heterosis. They defined the augmented additive and dominance effects in the RIL-based NCIII design: the former measures the net contribution of the QTL to the parental difference and the latter measures the net contribution of the QTL to midparent heterosis (MPH).…”