Data & code for - Hybridization may promote variation in cognitive phenotypes in experimental guppy hybrids

To test if hybridization can promote phenotypic variation in cognitive abilities, we compare the performance of guppies (Poecilia reticulata), Endler’s guppies (Poecilia wingei), and their experimental hybrids in a colour association and reversal learning tasks.In addition to analysing this data using standard cognition metrics (e.g. trials to learning criterion), we introduce a new approach to analyse and compare multidimensional cognitive phenotypes. We used Kernel Density Estimation (KDE) together with a geometric approach to quantify patterns of hybrid phenotypes and quantitatively compare phenotypic dispersion between hybrids and parentals and quantify the extent to which hybrids are transgressive and/or deviate from parental mean phenotypes. <br>This depository contains the datasheet that identifies the crossing group of each fish, the raw data for the associative learning and reversal learning tasks, and a processed data table containing information on the performance of the fish in both tasks (learning criterion). We also provide the R code used for all analyses and to generate the graphs presented in the manuscript.

为验证杂交能否促进认知能力的表型变异，我们针对孔雀鱼（Poecilia reticulata）、恩德勒孔雀鱼（Poecilia wingei）及其实验杂交后代，开展颜色联想学习与反转学习任务，并对比三者的行为表现。除采用标准认知指标（如达到学习标准所需的试验次数）对该数据集进行分析外，我们还引入了一种全新方法，用于分析并对比多维认知表型。我们将核密度估计（Kernel Density Estimation, KDE）与几何分析方法相结合，以量化杂交后代的表型分布模式，定量比较杂交后代与亲本群体间的表型离散程度，并量化杂交后代呈现超亲性状以及偏离亲本平均表型的程度。
本数据集仓库包含以下内容：用于标识每条实验鱼所属杂交组合的信息表、联想学习与反转学习任务的原始数据，以及包含两种任务中实验鱼表现信息（学习达标情况）的预处理数据表。此外，我们还提供了本研究所有分析及论文配图生成所用的R语言代码。