Prediction and analysis of antibody amyloidogenesis from sequences

Abstract

Antibody amyloidogenesis is the aggregation of soluble proteins into amyloid fibrils that is one of major causes of the failures of humanized antibodies. The prediction and prevention of antibody amyloidogenesis are helpful for restoring and enhancing therapeutic effects. Due to a large number of possible germlines, the existing method is not practical to predict sequences of novel germlines, which establishes individual models for each known germline. This study proposes a first automatic and across-germline prediction method (named AbAmyloid) capable of predicting antibody amyloidogenesis from sequences. Since the amyloidogenesis is determined by a whole sequence of an antibody rather than germline-dependent properties such as mutated residues, this study assess three types of germline-independent sequence features (amino acid composition, dipeptide composition and physicochemical properties). AbAmyloid using a Random Forests classifier with dipeptide composition performs well on a data set of 12 germlines. The within- and across-germline prediction accuracies are 83.10% and 83.33% using Jackknife tests, respectively, and the novel-germline prediction accuracy using a leave-one-germline-out test is 72.22%. A thorough analysis of sequence features is conducted to identify informative properties for further providing insights to antibody amyloidogenesis. Some identified informative physicochemical properties are amphiphilicity, hydrophobicity, reverse turn, helical structure, isoelectric point, net charge, mutability, coil, turn, linker, nuclear protein, etc. Additionally, the numbers of ubiquitylation sites in amyloidogenic and non-amyloidogenic antibodies are found to be significantly different. It reveals that antibodies less likely to be ubiquitylated tend to be amyloidogenic. The method AbAmyloid capable of automatically predicting antibody amyloidogenesis of novel germlines is implemented as a publicly available web server at http://iclab.life.nctu.edu.tw/abamyloid.

Figure 1. The receiver operator characteristic (ROC) curves of three general methods.

Figure 2. Three evaluation methods for AbAmyloid.

(A) There are 12 individual germline models. Each model is evaluated using a Jackknife test. (B) Only one model is constructed using a dataset of 12 germline (AA-432). (C) The leave-one-germline-out test is applied to evaluate the novel-germline prediction of AbAmyloid where each dataset of one germline is served as the test dataset of novel germline in turn.

Figure 3. Histogram and percentages for sequence pairs with sequence identities between training and test datasets.

Figure 4. Histogram for sequence pairs with sequence identities between amyloidogenic and non-amyloidogenic sequences.

Figure 5. Learning curves using various numbers of germlines for training classifiers.

The prediction performance is evaluated by using a leave-one-germline-out test.

Figure 6. Feature importance of Amino Acid Composition (AAC).

The feature with the largest value of mean decrease of Gini index (MDGI) is the most important.

Figure 7. Feature importance of Dipeptide Composition (DPC).

The feature with the largest value of mean decrease of Gini index (MDGI) is the most important.

Figure 8. The heatmap of DPC feature importance.

Figure 9. Feature importance of Physicochemical Properties (PPs).

The feature with the largest value of mean decrease of Gini index (MDGI) is the most important.

Figure 10. The 100% stacked column chart of the numbers of lysines (K) and putative ubiquitylated lysines (Ub-K).

Figure 11. Distribution of lysines for amyloidogenic and non-amyloidogenic antibodies.

Figure 12. Feature importance of 12 ubiquitylation features.

The feature with the largest value of mean decrease of Gini index (MDGI) is the most important.

Figure 13. Feature importance of 12 ubiquitylation features and Dipeptide Composition (DPC).

The feature with the largest value of mean decrease of Gini index (MDGI) is the most important.