Table 4. Rule summary for the whole training set when applying the simplified machine learning approach consisting of human-understandable attributes (ML-simple) for prediction of IVIG binding.
| Attributea | Low\highb | Classified correctlyc |
| Aromaticity | High | 53.8% |
| Polarity | Low | 27.7% |
| Frequency of tyrosine | High | 26.2% |
| Hydrophobicity | Low | 22.5% |
| Frequency of arginine | High | 19.7% |
| Summary factor 2 | High | 16.7% |
| Acidity | Low | 11.4% |
| Preference for β-sheets | Low | 4.3% |
| Summary factor 5 | High | 3.0% |
a
Details on the attributes, including the two summary factors “Summary Factor 2” and “Summary Factor 5” that combine 494 amino acid properties, are given in File S3 (2. Attributes for Machine Learning).
b
Whether the rules state the value of the attribute should be high or low for a peptide to be binding.
c
The percentage of binding peptides that were correctly classified by rules containing the attribute. This percentage roughly corresponds to the importance of the attribute.