Ecotoxicogenomics: interlaboratory comparability of microarray data [LabC_Ex]

Additional toxicity testing and regulation are needed as the number of compounds chronically released into aquatic coastal environments increases. Decision makers require comprehensive and efficient screening tools to detect adverse biological effects and to identify their cause. Transcriptomic analysis can complement traditional ecotoxicology data by providing mechanistic insight, identifying sub-lethal organismal responses and identifying contaminant classes underlying observed toxicity. Before transcriptomic information is used in monitoring and risk assessment, it is necessary to determine its reproducibility and detect key steps that impact the reliable identification of differentially expressed genes. A custom 15K-probe microarray was used to conduct transcriptomics analyses across six laboratories. Estuarine amphipods were exposed to cyfluthrin-spiked or control sediments for 10 days. Two sample types were generated: the first consisted of total RNA extracts (Ex) from exposed and control samples (extracted by one laboratory) and the second type consisted of exposed and control whole body amphipods (WB) from which RNA was extracted by each laboratory. A subset of genes were consistently identified as differentially expressed across all laboratories and sample types. Genes with the highest magnitude of differential expression (> 2 fold) were more likely to be consistently identified as differentially expressed across laboratories. Differentially expressed data had a higher degree of concordance across all laboratories (W = 0.7) in samples with similar RNA quality (Ex) when compared to WB samples (W = 0.5). Our results revealed that several factors can affect data comparability including RNA sample preparation, labeling, instrumentation and technical expertise. First dataset: RNA extracts (Ex) derived from whole bodies of 8 controls and 8 cyfluthrin-exposed amphipods were prepared by one single lab and analyzed (i.e., labeling, microarray hybridization and feature extraction) by all 6 laboratories. Second dataset: Each of the 6 labs received 8 control and 8 cyfluthrin-exposed whole body samples (WB) for RNA extraction and microarrays. Lab C was unable to analyze the WB samples and only submitted data for the Ex samples.

随着持续排入沿海水生环境的化合物数量持续攀升，亟需开展更多毒性测试与监管工作。决策者亟需兼具全面性与高效性的筛选工具，以检测有害生物效应并明确其诱因。转录组学（Transcriptomic）分析可弥补传统生态毒理学（ecotoxicology）数据的短板，通过提供机制层面的洞察、识别亚致死级生物体反应，以及揭示引发观测到的毒性效应的污染物类别，为生态毒理学研究提供补充。在将转录组学信息应用于环境监测与风险评估前，有必要明确其可重复性，并找出影响差异表达基因（differentially expressed genes）可靠鉴定的关键环节。
本研究采用定制化15K探针微阵列（custom 15K-probe microarray），开展跨6个实验室的转录组学分析。将河口钩虾暴露于添加氟氯氰菊酯（cyfluthrin）的沉积物或对照沉积物中，暴露时长为10天。本次实验生成两类样本：第一类为暴露组与对照组的总RNA提取物（以下简称Ex样本），由单一实验室完成提取；第二类为暴露组与对照组的完整钩虾体（以下简称WB样本），各实验室均对该类样本开展RNA提取操作。
研究发现，在所有实验室与样本类型中，均有部分基因被一致鉴定为差异表达基因。差异表达幅度超过2倍的基因，在不同实验室间被一致鉴定为差异表达基因的概率显著更高。与全虫体样本（WB样本，组内一致性W=0.5）相比，RNA质量相近的总RNA提取物样本（Ex样本）中，差异表达数据在各实验室间的一致性程度更高（W=0.7）。本研究结果表明，多项因素可影响数据可比性，包括RNA样本制备流程、标记操作、实验仪器配置与技术操作经验等。
首个数据集：由单一实验室制备8个对照组与8个氟氯氰菊酯暴露组钩虾的总RNA提取物（Ex样本），随后由全部6个实验室分别完成标记、微阵列杂交与特征提取等分析流程。
第二个数据集：6个实验室各收到8个对照组与8个氟氯氰菊酯暴露组的全虫体样本（WB样本），用于RNA提取与微阵列实验。其中实验室C无法完成WB样本的分析，仅提交了Ex样本的相关数据。