New Proposals to Estimate Unbiased Selection Gain and Coefficient of Variation in Traits Evaluated Using Score Scales

Abstract

937 RESEARCHR andom variables can be classified as qualitative, discrete quantitative, or continuous quantitative, whose scales of measurement can be nominal, ordinal, interval, or ratio (Resende et al., 2014). The nominal scale classifies the data into distinct categories, assigning names and arbitrary numerical correspondence to the variable's categories (e.g., gender: 1 = female and 2 = male). The ordinal scale classifies data into distinct categories, and in this case, the order or position is assigned to the categories (e.g., water quality level: 1 = good, 2 = intermediate, and 3 = bad). Either nominal or ordinal scales show no relation of difference or ratio between the numerical values of the scale (Levine et al., 2014;Resende et al., 2014). The interval scale is an ordered scale in which the difference between any two numbers in the scale is known; however, 0 is included as an arbitrary point (e.g., temperature in °C). Finally, a ratio scale is an ordered scale in which the distances between any two numbers in the scale are known, and their measurements include the true zero as the point of origin (e.g., height in centimeters) (Levine et al., 2014;Resende et al., 2014).In plant evaluations, several multicategorical variables are visually evaluated by score scales, such as plant architecture and grain appearance in common bean (Phaseolus vulgaris L.; Batista et al., 2017); lodging in common bean (Soltani et al., 2016), soybean [Glycine max (L.) Merr.; Akpertey et al., 2018], and wheat (Triticum aestivum L.; Iqbal et al., 2016); vigor in chickpea (Cicer arietinum L.; Sivasakthi et al., 2018); leaf senescence and leaf rolling in maize (Zea mays L.; Soni et al., 2018); herbicide tolerance in lentils (Lens culinaris Medik.; Sharma et al., 2018); salt tolerance in soybean (Do et al., 2018); injury caused by pests in cotton (Gossypium hirsutum ABSTRACTIn plant breeding, several multicategorical variables are evaluated using score scales that are treated as interval scales. However, statistics that use ratio operations, such as the experimental CV and the selection gain, are only appropriate for ratio scales data. Thus, this work aimed to propose strategies to mitigate the inconveniences faced in the data analysis of score scales, especially those involving CV and selection gain obtained from different scales. This work proposes the following strategies: (i) the use of a standard score scale with the properties of a ratio scale (ascending scale with 0 as the point of origin); (ii) the conversion of the scores from different scales into a common ratio scale; (iii) the adjustment of the data from an interval scale to a ratio scale using the Delta Scale (Scale d ui ), and (iv) the use of unbiased estimator for CV and selection gain in score scales data. The proposed strategies resulted in unbiased estimates of CV and selection gain from data of different score scales. These strategies have the potential to be used in the meta-analysis of data within and between plant breeding programs.

Potential parents of carioca bean for use in breeding aiming resistance to Colletotrichum lindemuthianum

“…These values ​​are similar to those observed in studies evaluating the resistance of common bean lines to anthracnose, under artificial inoculation conditions, when corrected according to Dos Anjos et al. (2019). High accuracy values ​​were observed for the severity of anthracnose races (Table 3).…”

Multipopulation recurrent selection: An approach with generation and population effects in selection of self‐pollinated progenies

“…Then the P × G effects were fitted as random. Coefficient of residual variance was estimated according to Anjos, Marçal, Carneiro, and Carneiro (2019) for score based traits (plant architecture, angular leaf spot severity, and grain appearance).…”

“…An unbiased estimator of genetic gain was used according to Anjos et al. (2019) for scores based traits such as plant architecture, angular leaf spot severity, and grain appearance.…”

“…We estimated genetic gains considering different selection intensities (best 30, 20, 10, and 1) and compared with those gains obtained when selection was made directly for the trait. An unbiased estimator of genetic gain was used according to Anjos et al (2019) for scores based traits such as plant architecture, angular leaf spot severity, and grain appearance.…”

Multi-Dimensional Machine Learning Approaches for Fruit Shape Recognition and Phenotyping in Strawberry

“…Descriptive categories for planar shapes (e.g., rhombic and reniform) have long played a role in plant systematics [17]. Categories may be either nominal [11, 18, 19], existing in name only, or ordinal, referring to a position in an ordered series or on a gradient [15, 16, 19]. Classification is often labor-intensive and prone to human bias, which can increase with task complexity and time requirement [20, 21].…”