Generative Modeling, Design and Analysis of Spider Silk Protein Sequences for Enhanced Mechanical Properties

Spider silks are remarkable materials characterized by superb mechanical properties such as strength, extensibility and light weight features. Yet, to date, limited computational models are available to fully explore sequence-property relationships for analysis and design. We propose a generative modeling, design and analysis technique applied to create novel spider silk protein sequences for enhanced mechanical properties. The model, pretrained on a large set of protein sequences, is fine-tuned on 1,033 major ampullate spidroin (MaSp) sequences for which associated fiber-level mechanical properties were measured. This process represents an end-to-end forward and inverse generative strategy. The dataset is established by curating a published silkome dataset. Performance is assessed through: (1) novelty analysis and protein type classification for generated spidroin sequences through BLAST searches, (2) property evaluation and comparison with similar sequences, (3) comparison of molecular structures, and (4) detailed sequence motif analyses. We generate silk sequences with property combinations that do not exist in nature and develop a deep understanding of the mechanistic roles of sequence patterns in achieving overarching key mechanical properties (e.g., elastic modulus, strength, toughness, failure strain). Other research that will be discussed includes applying this approach to expand the silkome dataset by using a combination of generative modeling and molecular simulations, thereby facilitating further sequence-structure analyses of silks, and establishing a foundation for synthetic silk design and optimization.