Original title:
Využití latentní sémantické analýzy pro virtuální screening
Translated title:
Utilization of latent semantic analysis in virtual screening
Authors:
Kolář, Jiří ; Hoksza, David (advisor) ; Škoda, Petr (referee) Document type: Bachelor's theses
Year:
2017
Language:
eng Abstract:
Title: Utilization of latent semantic analysis in virtual screening Author: Jiří Kolář Department: Department of Software Engineering Supervisor: RNDr. David Hoksza, Ph.D., Department of Software Engineering Abstract: Aim of this thesis is to investigate utilisation of latent semantic in- dexing in Virtual screening. We have examined existing VS method called lat- ent semantic structural indexing (LaSSI) and compared performance of different structural fingerprints. Additionally, we have developed a new model that com- pare fragments of molecules by usage of latent semantic indexing. Fragments are characterized by formula based counts and descriptors describing the physi- cochemical properties. Results of our methods are compared to VS techniques using directly standard fingerprints. Keywords: virtual screening cheminformatics ligand-based fingerprints ECFP TT latent semantic analysis LaSSI iii
Keywords:
cheminformatics; latent semantic analysis; virtual screening; cheminformatika; latentní sémantická analýza; virtuální screening
Institution: Charles University Faculties (theses)
(web)
Document availability information: Available in the Charles University Digital Repository. Original record: http://hdl.handle.net/20.500.11956/90456