Base Swap Mining: Discovery and Rational Optimization of Cryptic Human Proteins, with P44-G2 as a pLDDT-Validated R28G Variant of P44 from FGF21 +1 Frame

This work introduces Base Swap Mining, a computational method for the rational discovery and optimization of cryptic proteins encoded in alternative reading frames of the human genome. We first applied Cryptic Frame Mining to human FGF21, identifying P44, a 44- amino acid protein in the +1 frame [10.5281/zenodo.20398834]. We then applied Base Swap Mining using an A→G substitution rule to critical codons of P44. This yielded P44-G2, an R28G variant. AlphaFold2 structure prediction of P44-G2 shows a mean pLDDT of 86.1 with a stable coiled-coil architecture and a flexible hinge at Gly28, compared to pLDDT 75 for wild-type P44. The reduced positive charge and introduced flexibility predict enhanced solubility and reduced aggregation propensity. P44-G2 serves as proof-of-concept that Base Swap Mining can convert cryptic frame discoveries into drug-like candidates. All structural data, PDB files, and validation plots are included. Computational assistance provided by Meta AI.