The controlled design of bimetallic nanoparticles (BNPs) is a key goal in tailoring their catalytic properties. Recently, biomimetic pathways demonstrated potent control over the distribution of different metals within BNPs, but a direct understanding of the peptide effect on the compositional distribution at the interparticle and intraparticle levels remains lacking. We synthesized two sets of PtAu systems with two peptides and correlated their structure, composition, and distributions with the catalytic activity. Structural and compositional analyses were performed by a combined machine learning-assisted refinement of X-ray absorption spectra and Z-contrast measurements by scanning transmission electron microscopy. The difference in the catalytic activities between nanoparticles synthesized with different peptides was attributed to the details of interparticle distribution of Pt and Au across these markedly heterogeneous systems, comprising Pt-rich, Au-rich, and Au core/Pt shell nanoparticles. The total amount of Pt in the shells of the BNPs was proposed to be the key catalytic activity descriptor. This approach can be extended to other systems of metals and peptides to facilitate the targeted design of catalysts with the desired activity.