Figure 1L: Epitope maps and confusion matrix with BM5 Ab sequence inputs

One of AbEMap's applications is to map epitope on a given antigen crystal structure if only the sequence of the antibody is entered by the user. The results in this dataset are pdb files of the 40 antigens in Zhiping Weng's docking benchmark v. 5.0 with their b-factor changed to our protocol's score to reflect the likelihood of each atom being in the epitope when in complex with the respective antibody. The summary confusion matrix (residue based) for the epitope mapping done by AbEMap on the 40 antigens in BM5 with antibody sequences as the input is also included. The text file contains the True Positive (TP), False Positive (FP), False Negative (FN) and True Negative (TN) counts for each antibody-antigen complex at different residue ranking thresholds (Top1, Top5, Top10, ... Top120). This confusion matrix was used to generate Fig 1L.

AbEMap的应用之一为：当用户仅输入抗体序列时，可在给定的抗原晶体结构上完成表位（epitope）映射。本数据集的结果包含Zhiping Weng对接基准数据集v5.0中40种抗原的PDB文件，这些文件的B因子已被替换为我们的算法流程的评分，用以反映各原子在与对应抗体形成复合物时处于表位的可能性。本数据集还包含AbEMap以抗体序列为输入，对BM5中的40种抗原进行表位映射后得到的基于残基的汇总混淆矩阵。该文本文件包含了在不同残基排序阈值（Top1、Top5、Top10……Top120）下，每个抗体-抗原复合物的真阳性（TP）、假阳性（FP）、假阴性（FN）与真阴性（TN）计数。该混淆矩阵被用于生成图1L。