While computational protein design is capable of generating a large number of sequences for a specific fold, the practicalities of the current methods for testing the designs (cloning, expression, purification) become overwhelming (Johanson et al, 2016). In order to overcome these challenges, we have produced a truncated Green Fluorescent Protein (tGFP), that lacks strand 10 in its beta-barrel (Do and Boxer, 2011), resulting in significant loss of fluorescence compared to the mother protein. Addition of a synthetic peptide strand 10 (s10) results in complete recovery of the fluorescence of the full-length protein. By immobilizing the s10 on a solid support in a peptide array setting, we can study in massive throughput how different variants of s10 are capable of recovering the function of GFP. Preliminary data indicate that peptide linkers at the array’s surface are influencing the fluorescent signal, with charged residues being promising candidates for increasing the signal and dynamic range of the assay. The s10 substitution analysis with 7 of the 20 possible amino acids shows that the assay is capable of offering a detailed map of the relationship between the function of GFP (i.e. fluorescence) and the amino acid sequence of its beta-strand 10. We are on our way to establishing a novel platform for protein design that, in high-throughput and with rapid turnover, can test computational methods experimentally.