Under the Guidance of PI: Dr UCA JALEEL (IISc Research Unit, Bangalore) Swati Gandhi [Shah] 3.2 TCOF Fellow (MSc Bioinformatics, The Maharaja Sayajirao University of Baroda). TCOF 3 :Repositioning of Chemical compounds From Different Classes as part of Virtual Screening.
Under the Guidance of PI: Dr UCA JALEEL
(IISc Research Unit, Bangalore)
Swati Gandhi [Shah]
3.2 TCOF Fellow
(MSc Bioinformatics, The Maharaja Sayajirao University of Baroda)
TCOF 3 :Repositioning of Chemical compounds From Different Classes as part of Virtual Screening
Repositioning of Chemical compound database divided under three sub classes:-
-> Me and My Group worked on Pesticides showing anti TB activity:
As per the Flow Chart of the Previous Slide We have Initialized the WEKA Part; The algorithms are applied directly to a dataset; Training Set and Test Set generated on a ratio of 80:20
WEKA Model is generated with the Help of this Training and Test Set and Next Slide Defines the Step.
Module – Work Flow
Accessing the HTS bioassay data
All compounds sdf file
Upload the sdf file
Generate descriptor file
Open the CSV file in Excel
Append the bioassay result corresponding to the compounds
Bioassay result (all)
Select the actives and inactive compounds
Remove the useless attributes
TP %, FP<20%, Accuracy >70%
Apply classifier algorithms
Selection of best classifier model
Current Stage of Project is Tuning of Model Generated by WEKA:
We are trying to Tune the Model to the Most Stable state Applying the Cost Matrix on it .
We have generated the Results using different Classifiers like Naïve bayes and Random Forest We are trying to Tune the Model giving the Cost Matrix to it. Next Slide will draw some light on this.
Next Stage is to Go for Screening and then We will proceed Further
Sheet Defining the Results After Applying the Cost Matrix
1) Schierz AC. Virtual screening of bioassay data. J Cheminform. 2009 Dec 22;1:21. doi: 10.1186/1758-2946-1-21. PubMed PMID: 20150999.
2) Periwal V etal., Predictive models for anti-tubercular molecules using machine learning on high-throughput biological screening datasets. BMC Res Notes. 2011 Nov18;4:504. doi: 10.1186/1756-0500-4-504. PubMed PMID: 22099929.
3) Ekins S, etal., Combining Computational Methods for Hit to Lead Optimization in Mycobacterium Tuberculosis Drug Discovery. Pharm Res. 2013 Oct 17. [Epub ahead of print] PubMed PMID: 24132686.
4) Enviornmental Protection Agency
Heartiest Thanks and Acknowledgement: 1) Prof. Dr. Samir Bramachari2)Dr Jaleel (PI TCOF3)3) Dr Bheemarao Ugarkar4)OSDD Team5)IISc Research Unit, Bangalore6) Group Members [Yatindra Yadav and Ayisha Safeeda]