Feedforward Neural Networks for investigating the Conformational Changes of Macrobiomolecules
Feedforward Neural Networks for investigating the Conformational Changes of Macrobiomolecules
Golnaz Ansarihadipour,1,*Ali Hakimzadeh,2Hadi Ansarihadipour,3
1. Veterinary Faculty, Islamic Azad University, Karaj Branch, Karaj, Iran 2. Payam Noor University of Tehran, East Tehran Branch. 3. Department of Biochemistry and Genetics, School of Medicine, Arak University of Medical Sciences, Arak, Iran
Introduction: The future of medicine will certainly involve the integration of artificial intelligence into laboratory methods to improve the accuracy and reliability of experiment data.
Methods: Artificial neural networks (ANNs) are applied in various disciplines such as: cytomorphology, immunohistology, cell differentiation, morphological features, automated flow cytometry, chromosome banding analysis, chromosome classification, analysis of molecular profiles, identification of therapeutic candidates and drug discovery. Moreover, ANNs can make accurate predictions in diseases with complex conditions. Hemoglobin is highly susceptible to conformational changes due to several different factors including hemolysis, ineffective erythropoiesis, iron overload, inflammation and increased production of reactive oxygen species.
Results: ANNs can be trained on datasets containing information about hemoglobin modifications, changes in its absorbance spectrum and its reaction kinetics with specific oxidants. By analyzing this data, ANNs can learn to identify specific patterns or relationships between the modifications of Hb and various biological and pathological outcomes. Our presentation will demonstrate how ANNs can fruitfully develop new methods for studying the structural changes of Hb in healthy and disease conditions.
Conclusion: At the present article, we first discuss about basic concepts of ANNs and focus on bringing this mathematical framework closer to medicine. Then, we introduce the modifications of hemoglobin which can be studied by spectrophotometric analysis at different wavelengths. Finally, we detail how to customize the analysis, structure, and learning of ANNs to better address the conformational changes of hemoglobin.