A sequence-dependent classification algorithm for Crohn’s Disease – causing NOD2 protein mutations


  • Jose Isagani B. Janairo De La Salle University
  • Marianne Linley L. Sy-Janairo St. Luke’s Medical Center – Global City




Artificial neural networks, Inflammatory bowel disease, Machine learning, Personalized medicine


Certain NOD2 protein mutations have been associated with the onset of the inflammatory bowel disease, Crohn’s Disease (CD). NOD2 is involved in the inflammatory response of the gut to the microbial community, wherein its functional impairment through mutations may lead to CD progression. Considering the significant role that NOD2 plays in CD pathogenesis, predicting whether a specific type of NOD2 mutation is the cause of CD can greatly aid the accuracy of the disease diagnosis. Hence, a novel sequence-based classification algorithm built on artificial neural network (ANN) is herein presented that can predict whether a specific NOD2 mutation can cause CD or not. The NOD2 mutant types and their association with CD were taken from literature, and the calculated sequence-order coupling numbers were used as the classification predictors. The formulated ANN classifier exhibited satisfactory predictive ability, with 82.4 % accuracy, 62.5 % sensitivity, 100 % specificity, 100 % positive predictive value, and 75 % negative predictive value. The presented ANN classifier provides a proof-of-concept that predicting the onset of CD from NOD2 protein variant is possible.

Author Biographies

Jose Isagani B. Janairo, De La Salle University

Biology Department, De La Salle University, 2401 Taft Avenue, Manila 0922, Philippines

Marianne Linley L. Sy-Janairo, St. Luke’s Medical Center – Global City

Institute of Digestive and Liver Diseases, St. Luke’s Medical Center – Global City, Rizal Drive, Taguig 1634, Philippines




How to Cite

Janairo, J. I. B. ., & Sy-Janairo, M. L. L. . (2020). A sequence-dependent classification algorithm for Crohn’s Disease – causing NOD2 protein mutations . Nova Biotechnologica Et Chimica, 19(1), 52–60. https://doi.org/10.36547/nbc.v19i1.577



Research Articles