Amyloid propensity prediction neural network (APPNN) is an amyloidogenicity propensity predictor based on a machine learning approach through recursive feature selection and feed-forward neural networks, taking advantage of newly published sequences with experimental, in vitro, evidence of amyloid formation. This approach relies on the assumptions that, i) small peptide stretches within an amyloidogenic protein can act as amyloid forming facilitators that will eventually direct the refolding of the protein along a path involving the formation of an energetically favourable amyloid conformation; ii) the minimum length of these facilitator sequences or hot spots comprises six amino acids; iii) the amyloidogenicity propensity value per amino acid corresponds to the highest value obtained from all six amino acid windows that contain that amino acid; and iv) a peptide or protein is considered amyloidogenic if at least one stretch or hot spot is found within the sequence.
Carlos Família, Sarah R. Dennison, Alexandre Quintas, David A. Phoenix
Maintainer: Carlos Família <carlosfamilia@gmail.com>
| Package: | appnn |
| Type: | Package |
| Version: | 1.0-1 |
| Date: | 2024-12-05 |
| License: | GPL-3 |
The amyloidogenic propensity prediction neural network is composed by three functions, the function appnn which performs the propensity prediction calculations, the function print that prints to the console the prediction results, and function plot that generate plots of the prediction results.
Família, C., Dennison, S. R., Quintas, A. & Phoenix, D. A. Prediction of Peptide and Protein Propensity for Amyloid Formation. Plos One 10, e0134679 (2015).
sequences <- c('STVIIE','KKSSTT','KYSTVI')
predictions <- appnn(sequences)
print(predictions)
plot(predictions,c(1,2,3))
Run the code above in your browser using DataLab