The present study aimed to compare the accuracy and reliability of four standard methods used for classification of people as taster or non-tasters based on their sensitivity to PROP (6-n-propylthiouracil). A panel consisting of 21 subjects was tested for threshold and suprathreshold sensitivity of sodium chloride, PROP, and genotyped for TAS2R38. Two threshold methods, staircase and modified Harris–Kalmus, were used to obtain detection and recognition thresholds and compared for accuracy and repeatability. Similarly, two suprathreshold techniques, the just noticeable differences (JND) and the general labeled magnitude scale (gLMS), were used to determine Weber fractions and individual psychophysical functions and compared for accuracy and repeatability. Results show both threshold methods have been able to correctly separate people into two groups of tasters and non-tasters, with the staircase method having a lower variability among subjects. On the suprathreshold front, we found differences in sensitivity between tasters and non-tasters when comparing Weber fractions and psychophysical functions; however, our data suggest that clustering people without previous knowledge of their taster status is less accurate when using Weber fractions. Intensity ratings are more reliable to classify people into tasters and non-tasters. Results show that the staircase for threshold measurement and the gLMS methods are more reliable methods than Harris–Kalmus and JND for phenotyping people and can be used in large-scale studies in the quest to discover new genotype–phenotype associations.

中文翻译：

---

用于味觉表型的阈值和超阈值方法的可靠性：用 PROP 和氯化钠表征。

本研究旨在比较四种标准方法的准确性和可靠性，这些方法用于根据人们对 PROP（6- n-丙基硫氧嘧啶）的敏感性将人分类为品尝者或非品尝者。由 21 名受试者组成的小组测试了氯化钠、PROP 的阈值和超阈值敏感性，并对TAS2R38 进行了基因分型. 两种阈值方法，staircase 和修改的 Harris-Kalmus，用于获得检测和识别阈值，并比较准确性和可重复性。类似地，两种超阈值技术，即显着差异 (JND) 和一般标记量级 (gLMS)，用于确定韦伯分数和个体心理生理功能，并比较准确性和可重复性。结果表明，两种阈值方法都能够将人们正确地分为品尝者和非品尝者两组，阶梯法在受试者之间的变异性较低。在超阈值方面，我们发现在比较韦伯分数和心理生理功能时，品尝者和非品尝者的敏感性存在差异；然而，我们的数据表明，在使用 Weber 分数时，对先前不了解其品尝者状态的人进行聚类不太准确。强度等级更可靠地将人们分为品尝者和非品尝者。结果表明，与 Harris-Kalmus 和 JND 相比，阈值测量阶梯法和 gLMS 方法是更可靠的方法，可用于大规模研究以发现新的基因型 - 表型关联。