Login / Signup

Predicting the stability of homotrimeric and heterotrimeric collagen helices.

Douglas R WalkerSarah A H HulganCaroline M PetersonI-Che LiKevin J GonzalezJeffrey D Hartgerink
Published in: Nature chemistry (2021)
Robust methods for predicting thermal stabilities of collagen triple helices are critical for understanding natural structure and stability in the collagen family of proteins and also for designing synthetic peptides mimicking these essential proteins. In this work, we determine the relative stability imparted on the collagen triple helix by single amino acids and interactions between amino acid pairs. Using this analysis, we create a comprehensive algorithm, SCEPTTr, for predicting melting temperatures of synthetic triple helices. Critically, our algorithm is compatible with every natural amino acid, can evaluate both homotrimers and heterotrimers, and accounts for all possible helix compositions and registers, including non-canonically staggered helices. We test and optimize our algorithm against 431 published collagen triple helices to demonstrate the quality of our predictive system. Finally, we use this algorithm to successfully guide the design of an ABC heterotrimer possessing high assembly specificity.
Keyphrases
  • amino acid
  • machine learning
  • deep learning
  • wound healing
  • tissue engineering
  • neural network
  • high resolution
  • randomized controlled trial
  • meta analyses