| |
|
| Technology > kNN-MFA |
|
|
|
|
|
|
kNN-MFA is a novel methodology. Unlike conventional QSAR regression methods, this methodology can handle non-linear relationships of molecular field descriptors with biological activity, thus making it a more accurate predictor of biological activity.
Conventional correlation methods try to generate linear relationship with the activity, whereas kNN is inherently non-linear method and is better able to explain activity trends.
kNN-MFA is built around the conceptually simple approach of pattern recognition working on active analog principle. The best part is that kNN-MFA has the ability to generate several models using user defined training and test set. |
| |
 |
| |
|
| GQSAR: A patent pending technology for fragment based QSAR developed by VLife that enhances use of QSAR for design optimization of molecule delivering highly specific site directed clues for design modification. |
| VLifeSCOPE: A novel technology application creating a hybrid approach for lead optimization and prioritization of design for a given purpose from a library of molecules. |
| LeadGrow+: An extension to the combinatorial library generation capability of VLifeMDS that significantly expands the chemical universe by enabling template substitution. |
| VLifeAutoQSAR: Unique automated approach to conduct QSAR that provides a best result based on a consensus of multiple QSAR models generated. |
| Aakar: A powerful and fast alignment independent shape search method with or without taking into consideration the chemical pharmacophoric features. |
| VLifeWorkFlow: A tool to customize and automate the discovery protocols of users using the CADD components. |
|
|
 |
|
Pub: J. Chem. Inf. Model. 46, 24-31,2006
Three-Dimensional QSAR Using the k-Nearest Neighbor Method and its Interpretation,
Subhash Ajmani, Kamalakar Jadhav and Sudhir A. Kulkarni |
|
|
|
kNN MFA does not assume a linear relation between activity and molecular properties and the inherent non-linearity in the method leads to improved models resulting in better predictive ability |
kNN MFA has an intrinsic approach of pattern recognition on active analog principle and its ability to exhaustively scan several possible models utilizing user defined selections on the data set |
The QSAR models derived with kNN MFA can lead to generation of a library of molecules that satisfy one or many design considerations suggested by kNN-MFA model |
| |
|
| |
|
 |
A novel regression technique for modeling non-linear activity and property data View |
 |
Application of enzymes to enhance drug action View |
 |
Target Identification for existing nutraceutical molecule View |
|
|
|
|
|
|
|
|
"This new QSAR methodology gives QSARpro, a decisive edge over conventional QSAR. The ability to combine kNN with MFA is a unique approach which I came across only in QSARpro from VLife. It is now a method of choice in my research."
Dr. S.P.Gupta
Ex-BITS, Pilani |
|
|
| GQSAR |
| For site specific design clues |
|
|
|
|
| Taking cognizance of non-linearity |
|
|
|
|
| Aakar |
| Shape Based Screening |
|
|
|
|
|
|
|
|
|
|
|
|
| LeadGrow |
| Combinatorial library generation |
|
|
|
|
|
|
|
|
