Journal: Protein Science : A Publication of the Protein Society

Article Title: Accurate affinity models for SH2 domains from peptide binding assays and free‐energy regression

doi: 10.1002/pro.70317

Figure Lengend Snippet: Overview of concerted experimental and computational strategy for generating SH2‐peptide binding free energy models. (a) Design of peptide‐display libraries. (b) Schematic showing how a randomized bacterial display library underwent repeated bead‐based affinity selection for SH2 binding. In each selection round, the library was sequenced before and after selection. (c) Overview of the regression framework used to learn energetic binding models from the sequencing data. For each possible binding site, the energy received independent additive contributions from the residues flanking the phosphorylated tyrosine, thus controlling for the binding‐site context wherein the residues reside. These energy contributions were estimated using maximum likelihood estimation, where the likelihood of the observed sequence counts was evaluated by first computing the total affinity for each observed sequence (controlling for multiple possible binding offsets and non‐specific binding) and then computing the binomial likelihood for each round, assuming linear section. (d) Sequence logo displaying the inferred energy contributions as letters whose height reflects the magnitude of the contributions, relative to the mean for each position.

Article Snippet: All peptides containing phosphorylated tyrosine, both fluorescently labeled and non‐labeled, were produced by SynPeptide (China).

Techniques: Binding Assay, Selection, Sequencing