Additional file 1: of modPDZpep: a web resource for structure based analysis of human PDZ-mediated interaction networks

All Supplementary tables are merged into single Excel file. Table S1. ROC- and PR-AUC values for 14 Human PDZ domains with artificial negative interaction data. Approximately equal number of binder and nonbinder were considered in small sets and average AUC is calculated for each domain. For SCRIBBLE1_LAP4_3, PSCDBP and NHERF2_2, equal number of non-binder peptides were randomly selected and AUC is calculated. This is repeated twice to get average AUC. Table S2. List of 43 PDZ domains included in modPDzpep testset. PDZ domains which cannot be predicted by POW are marked as ‘N’, while those used in the training are marked ‘T’. The list of test set peptides for these 43 PDZ domains is available under Benchmark link of modPDZpep server. Table S3. List of phage display peptides which showed exact match in human C-terminal proteome Table S4. Amino acid variants present on human PDZ domains. (XLS 55 kb)

所有补充表格已合并为单个Excel文件。
表S1：基于14个人类PDZ结构域（PDZ domain）人工负相互作用数据的ROC-AUC与PR-AUC值。在小型数据集中纳入数量大致相当的结合肽与非结合肽，并针对每个PDZ结构域计算平均AUC值。针对SCRIBBLE1_LAP4_3、PSCDBP及NHERF2_2三个结构域，随机选取等量的非结合肽段并计算AUC，该流程重复两次以获取平均AUC值。
表S2：modPDzpep测试集中包含的43个PDZ结构域列表。无法通过POW模型进行预测的PDZ结构域标记为"N"，用于模型训练的PDZ结构域标记为"T"。上述43个PDZ结构域对应的测试集肽段列表可从modPDZpep服务器的Benchmark（基准测试）链接处获取。
表S3：在人类C端蛋白质组中存在精确匹配序列的噬菌体展示肽段列表。
表S4：人类PDZ结构域上存在的氨基酸变异体。（XLS 55 kb）