Our specially designed library covers about 5 billion different, random 16-mer peptide sequences, covering most 6-mer amino acid combinations. Due to the special design and care in the generation, the peptide phage library does not contain overrepresented single sequences, even when the library is replicated. This is allowing the following straightforward approach, which will fail, when other starting libraries are used.

Only minimal amounts of antibodies (µg) or serum (ca 100 µl) are required. The selection and enrichment of binders requires theoretically not more than a few hundred antibody molecules, so even the epitopes of antibodies in a serum can be identified.

For each epitope up to thousand variants of 16-mer sequences resembling the epitope’s motif are enriched in a single step, but they are still hidden in up to one million different other sequences. But due to the statistically even amino acid distribution even a minimal enrichment is sufficient for the following steps.

NGS is error prone. With several specially developed algorithms, we can sort out sequencing errors. In our set up with illumina sequencers up to 30% of the obtained single sequences are expected to contain errors and are removed prior to the following analysis. This avoids unnecessary and expensive oversampling, which would multiply the project costs.

Data sets with up to a million peptides containing sometimes only a few hundred enriched sequences with epitope motifs are generated by translating the peptide genes. Enrichment will not be detected on the base of single sequences, but it is using the statistical analysis to identify a pool of similar sequences sharing enriched motifs.

Statistical fishing for epitope sequences out of these large data sets is done with the help of a proprietary software that allows to pick enriched continuous and discontinuous peptide motifs resembling the antigen or just sharing sequence features. Surrounding amino acids of the 16-mer peptides often support special peptide epitope conformations required for binding to the antibody.

Epitope fingerprints obtained combining the information from many peptides allow to understand not only the location of the epitope but also the individual preferences of antibodies with respect to surrounding or intermitting amino acid residues. Analysis of aligned sequences helps to predict specificity with respect to bound protein in comparison to related or non-related proteins from the host proteome or other species. This is unique since no additional experiments are initially required.