The Spider silk exhibits exceptional mechanical behavior among natural materials, characterized by high toughness, extensibility, and a remarkable sensitivity to environmental conditions such as humidity. A striking property of spider silks is the supercontraction - a significant shrinkage upon wetting - deeply rooted in the molecular structure of its proteins. Understanding how the primary sequences of silk proteins influence the macroscopic fiber behavior is essential for advancing biomimetic materials. In this study, we propose an enhancement to a previously proposed multiscale model by integrating novel insights from data-driven modelling [1]. Specifically, using Evolutionary Polynomial Regression (EPR), a machine learning technique that combines genetic programming with symbolic regression, we investigate the relationship between the molecular composition of silk’s main proteins — MaSp1 and MaSp2 — and the fiber’s response to hydration. Our analysis of recent multiscale experimental data across different silk types [2] identifies some key protein sequence motifs: the repeat length of the protein MaSp2 and the polyalanine regions of the protein MaSp1 emerge as fundamental determinants of hydrated silk fiber behavior. EPR generates interpretable mathematical relationships linking these sequence features to fiber contraction, which we interpret in mechanical terms. We propose that the polyalanine domains of MaSp1 regulate β-sheet misalignment, accommodating the shortening of softer regions during supercontraction. On the other hand, the repeat length of MaSp2 governs the cross-linking interactions that stabilize amorphous chains in the dry state, while hydration disrupts them and triggers macroscopic fiber contraction. Validation against experimental data from the Silkome database [2] confirms the predictive capability of the proposed model, bridging molecular-scale protein structure with fiber mechanics. This work elucidates the molecular mechanisms underlying spider silk’s response to environmental changes, and establishes a predictive framework for the design of silk-inspired materials with tunable mechanical properties.