Computational design of conformation-affecting (CB) mutations is a rapid method that uses contrast scoring by inverse conformational models to predict protein variants favoring desired conformational states.[1][2][3] Most natural proteins alternate between different conformational states associated with specific functions.[1] They validated the CB method on seven diverse datasets of deep mutational scanning.[1][2][3] They successfully predicted protein variants of K-Ras, SARS-CoV-2 spike, β2 adrenergic receptor and Src kinase with enhanced conformation-specific functions such as enhanced effector binding or enzyme activity.[1][2][3] When applied to the enzyme lipoic acid ligase (LplA), they revealed a hitherto unknown mechanism of conformational closure of sequence specificity.[1][2][3] Variants favoring the "open" conformation were highly promiscuous, while variants in the "closed" conformation were more specific than the wild type, improving the use of LplA for specific labeling of proteins with fluorophores in living cells.[1][2][3] CB is fast, simple and versatile, available on GitHub.[1][2][5]