NHLBI‐ AbDesigner , artificial intelligence‐based software for design of peptide‐directed antibodies

Although protein mass spectrometry is useful for discovery science, follow up from such studies often requires development of high quality antibodies. Here we describe a new software tool called NHLBI‐ AbDesigner written in Java that predicts optimal sequences to be used for production of synthetic peptides for immunization. The choice of the immunizing peptides is based on a need to optimize immunogenicity, antibody specificity, multi‐species conservation, and robustness in the face of post‐translational modifications (PTMs). Current protocols for doing this require multiple manual steps using different computational tools and elements of human judgment that require experience. The new software predicts optimal sites along the primary sequence of proteins by calculating and extracting information as follows: 1) immunogenicity score based on secondary structure prediction and hydropathy; 2) uniqueness score [predicting specificity of an antibody]; 3) conservation score [predicting likelihood of multi‐species recognition of an antibody]; and 4) protein features that describe regions or sites of interest such as PTMs. It displays the prediction in an easy‐to‐view graphic user interface including heat maps for the scored predictors. NHLBI AbDesigner is available in the form of stand‐alone and web applications. It can also perform batch processing for a large‐scale production of antibodies which can be exploited for antibody‐based proteomics tools.