This guide frames developability assessment as an early screen for antibody candidates that may be difficult to manufacture, formulate, or maintain. It surveys computational approaches intended to flag liabilities before costly experimental development.
Aggregation and solubility are treated as related but distinct properties. Sequence-based methods can rapidly screen large candidate sets, while structure-aware tools add information about exposed residues and molecular context; the article recommends interpreting more than one signal together.
The guide also covers viscosity and self-interaction, including surface charge patterns and learned sequence-based descriptors. These methods help rank potential risk, but predictions should be calibrated against measurements from the intended antibody format and concentration range.
Stability and chemical liabilities form another group of checks. Sequence motifs and structural features can point to risks such as oxidation, deamidation, unstable regions, or unintended modifications, while mutation and energy-estimation tools can support follow-up design decisions.
Immunogenicity assessment focuses on identifying candidate T-cell epitopes and considering how antibody sequence and format relate to human immune recognition. The article distinguishes these computational estimates from experimental immune-response measurements.
The guide concludes with a practical filtering approach: combine orthogonal predictions, prioritize candidates with acceptable profiles, and validate the ranking experimentally. Developability scores are used as relative risk indicators, not guarantees of clinical performance.