Intelligent Palate Architecture Platform
The Intelligent Palate Architecture Platform (IPAP) is Scoville Splice's internal computational design pipeline. It transforms a flavor target (expressed as an IFI vector) into a validated genomic construct through four sequential layers: genomic data ingestion, ESM2 mutation scoring, metabolic flux analysis, and IFI receptor scoring.
Layer 1: Genomic Data Ingestion
Fully sequenced reference genomes for Capsicum chinense, Capsicum annuum, and the target host organism are loaded. For each proposed construct, the platform maps target loci, identifies orthologous genes, and designs guide RNAs for any CRISPR editing steps.
Layer 2: ESM2 Mutation Scoring
Codon-optimized variants of each biosynthesis enzyme are scored using Meta's ESM2 protein language model. Variants with mutation scores below threshold are eliminated. Those that pass proceed to metabolic modeling.
Layer 3: Metabolic Flux Analysis
FBA models the host's genome-scale metabolic network with the new biosynthesis reactions added. It predicts yield range, competing pathway conflicts, and supplementary edit recommendations.
Layer 4: IFI Receptor Scoring
Predicted compound profiles are translated into IFI scores using receptor binding affinity data and cell-based assay calibrations. Constructs are validated against the eight-axis IFI at protoplast level before catalog listing.