Data supporting manuscript: "Transgenerational effects of heat shock on Drosophila melanogaster gene regulation and life history depend on adaptive environment"

[Description of methods used for collection/generation of data] A full description of the methodology is available in the materials and methods section of the associated manuscript ("Transgenerational effects of heat shock on Drosophila melanogaster gene regulation and life history depend on adaptive environment"), with extra detail provided in the supplementary material of the manuscript. We provide a very brief summary here. Experimental design: Drosophila melanogaster from two populations (Manzanares in Spain and Akaa in Finland) were brought to the lab, and lab populations were set up, using the offspring of 10 F0 females from each population. In the F2, the critical thermal limits (CTMax) of female flies from each population were tested using heat ramping experiments. Then in the F3 we looked at heat shock responses by comparing female flies subject to a heat stress (one hour at 37 C) with unheated controls. Immediately following heat shock, female ovaries were dissected for RNA-seq and ATAC-seq analysis (pooled samples of 30 individuals, triplicated for each treatment) to investigate the effects on the transcriptome and epigenome. The bodies of flies prepared for ATAC-seq were prepared for DNA WGS reseq in order to characterise the presence of transposable elements (TEs) in these populations. We quantified the effect of heat shock on life history traits by measuring the number of eggs produced by 24 heat shocked and 24 control females from each population, the number that developed into pupae and adults, the viability of the eggs, and the time taken to pupate and eclose. These traits were measured in 4 separate F4 cohorts based on eggs laid in a 48 hour period: 0-2 days, 2-4 days, 4-6 days and 12-14 days. We then maintained 96 lines (24 females per treatment, 2 populations) for a further 2 generations, and in the F6 generation we repeated our RNA-seq and ATAC-seq experiments to see if there were any detectable transgenerational effects of heat shock (from the F3) on the transcriptomes or epigenomes of their great-grand-offspring. We also repeated the measurement of life history in the F7 generation (three generations after the previous phenotypic measurements) to check for transgenerational effects on these traits. Bioinformatics and statistics: RNA-seq data was analysed using kallisto, ATAC-seq data was analysed using the nextflow nf-core/atac pipeline, and the two data sets were integrated using the R package intepareto. Reference TEs were characterised using Tlex3, while non-reference TEs were characterised using the consensus insertions discovered by two packages: Temp2 and PoPoolationTE2. TE results were integrated using custom command line scripts. Differential expression and differential accessibility were assessed using DESEq2, with set analysis used to compare results between populations and across generations. Linear regression was used to assess the relationship between expression and accessibility in the F3 and between expression in the F3 and the F6 among key sets of genes. Chi-square tests were used to test for associations between TEs and specific patterns of expression and accessibility, and functional enrichment of GO terms was tested for various sets of genes specifically those that were i) both differentially expressed and differentially accessible in the F3; ii) DEG in the F3 and F6, and iii) associated with TEs and directional change in expression or accessibility (in most cases, sets of genes were population specific). The effect of heat shock or ancestral heat shock on phenotypic traits was tested using a variety of statistical models: linear mixed effects models (lmer) for ctmax and time to eclosion and pupation, generalised linear models (glm) with binomial error distribution for viability data, and glm with poisson error distributions for egg, pupae and adult counts. In most cases cohort was included as a random effect, so most models were run using glmer.

[数据收集/生成方法描述]
本文仅提供简要概述。实验设计：将来自两个种群（西班牙的Manzanares种群和芬兰的Akaa种群）的黑腹果蝇（Drosophila melanogaster）引入实验室，并利用每个种群10只F0代雌性个体的后代建立实验室种群。在F2代中，通过热梯度实验测定了每个种群雌性果蝇的临界热极限（CTMax）。随后在F3代中，通过比较经受热应激（37℃处理1小时）的雌性果蝇与未受热处理的对照组，研究其热休克反应。热休克后立即解剖雌性卵巢，用于RNA-seq和ATAC-seq分析（每个处理组取30只个体的混合样本，进行三次重复），以探究对转录组（transcriptome）和表观基因组（epigenome）的影响。用于ATAC-seq的果蝇躯体样本同时用于DNA全基因组重测序（WGS reseq），以鉴定这些种群中转座元件（transposable elements，TEs）的存在情况。我们通过测定每个种群中24只热休克处理雌性和24只对照雌性的产卵数、发育为蛹和成虫的数量、卵存活率以及化蛹和羽化所需时间，量化了热休克对生活史性状的影响。这些性状在F4代的4个独立队列中进行测定，队列划分基于产卵的48小时时间段：0-2天、2-4天、4-6天和12-14天。随后，我们将96个品系（每个处理组24只雌性，2个种群）继续饲养2代，并在F6代重复RNA-seq和ATAC-seq实验，以观察F3代热休克是否对其曾孙代的转录组或表观基因组产生可检测的跨代效应。我们还在F7代（距前一次表型测定三代后）重复了生活史性状的测定，以验证这些性状是否存在跨代效应。
生物信息学与统计学分析：RNA-seq数据使用kallisto进行分析，ATAC-seq数据使用nextflow nf-core/atac流程进行分析，两个数据集通过R包intepareto整合。参考转座元件（TEs）使用Tlex3鉴定，而非参考转座元件通过两个软件包（Temp2和PoPoolationTE2）发现的共有插入序列进行鉴定。转座元件分析结果通过自定义命令行脚本整合。差异表达和差异可及性（differential accessibility）使用DESEq2进行评估，通过集合分析比较种群间和世代间的结果。采用线性回归分析评估F3代中基因表达与可及性的关系，以及关键基因集在F3代和F6代之间的表达关系。采用卡方检验验证转座元件与特定表达模式及可及性模式之间的关联，并对各类基因集（尤其是以下三类：i）F3代中同时存在差异表达和差异可及性的基因；ii）F3代和F6代中的差异表达基因（differentially expressed genes，DEG）；iii）与转座元件相关且表达或可及性存在定向变化的基因）进行基因本体（Gene Ontology，GO）术语功能富集分析。多数情况下，这些基因集具有种群特异性。采用多种统计模型检验热休克或祖先热休克对表型性状的影响：对于临界热极限（CTMax）、羽化及化蛹时间，使用线性混合效应模型（lmer）；对于存活率数据，使用具有二项误差分布的广义线性模型（glm）；对于卵、蛹和成虫数量，使用具有泊松误差分布的广义线性模型（glm）。多数情况下，队列被视为随机效应，因此大多数模型使用glmer运行。