Dear Sir/madam
Looking forward your suggestion regarding 3D QSAR analysis. I uses Discovery Studio 3.5. I have worked with 3DQSAR analysis of a metalloprotease. I sets the 3DQSAR protocol according to the following:
1. Retrieved inhibitors with Ki value (60 inhibitors)
2. Create Activ (Ki value) uncert (2) under data table
3. Generate test and training set (by random method/75%training set)
4. Feature mapping by using active molecule.(HD,HA,H,RA)
5. Run the protocol by using default parameter.
It shows test set cross validation (i.e-q2) of Hypo1: 0.113
Correlation coefficient and RMS of Hypo1: 0.9180 and 1.127 respectively.
Here the validation result is very bad.
How do I increase the q2 of test set approximately to 0.7 - 0.9.
It will be very helpful for me if you kindly suggest the proper protocol set up of 3D QSAR in Discovery Studio 3.5.
Ankur Chaudhuri
Senior Research Fellow
West Bengal State University
Kolkata-700126
India
Looking forward your suggestion regarding 3D QSAR analysis. I uses Discovery Studio 3.5. I have worked with 3DQSAR analysis of a metalloprotease. I sets the 3DQSAR protocol according to the following:
1. Retrieved inhibitors with Ki value (60 inhibitors)
2. Create Activ (Ki value) uncert (2) under data table
3. Generate test and training set (by random method/75%training set)
4. Feature mapping by using active molecule.(HD,HA,H,RA)
5. Run the protocol by using default parameter.
It shows test set cross validation (i.e-q2) of Hypo1: 0.113
Correlation coefficient and RMS of Hypo1: 0.9180 and 1.127 respectively.
Here the validation result is very bad.
How do I increase the q2 of test set approximately to 0.7 - 0.9.
It will be very helpful for me if you kindly suggest the proper protocol set up of 3D QSAR in Discovery Studio 3.5.
Ankur Chaudhuri
Senior Research Fellow
West Bengal State University
Kolkata-700126
India