RT Journal Article
SR Electronic
T1 Binding Affinity Regression Models from Repeats Mutation in Polyglutamine Disease
JF bioRxiv
FD Cold Spring Harbor Laboratory
SP 281949
DO 10.1101/281949
A1 P R Asha
A1 M S Vijaya
YR 2018
UL http://biorxiv.org/content/early/2018/03/14/281949.abstract
AB Diagnosing and curing neurodegenarative disorder such as spinocerebellar ataxia is complicated when there is differences in formation of protein sequences and structures. Affinity prediction plays vital role to identify drugs for various genetic disorders. Spinocerebellar ataxia occurs but mainly it occurs due to polyglutamine repeats. This research work aims in predicting the affinity of spinocerebellar ataxia from the protein complexes by extracting the well-defined descriptors. Regression models are built to predict the affinity through machine learning techniques coded in python using the Scikit-Learn framework. Energy complexes and protein sequence descriptors are defined and extracted from the complex and sequences. Results show that the SVR is found to predict the affinity with high accuracy of 98% for spinocerebellar ataxia. This paper also deliberates the results of statistical learning carried out with the same set of complexes with various regression techniques.