Lorentz JÄNTSCHI (lori) works ?id=99
- [id] => 99
- [recorddate] => 2007:01:14:16:36:48
- [lastupdate] => 2007:01:14:16:36:48
- [type] => conference
- [place] => Tale, Slovakia
- [subject] => chemistry - biochemistry; chemistry - computational; informatics - models implementation; mathematics - modeling; research - evaluation; research - methodology
- [relatedworks] =>
- 2 (average):
- Structure-activity relationships on the molecular descriptors family project at the end, ?id=130
- 4 (some):
- Mobile phase optimization in three solvents high performance thin-layer chromatography: methodology and evaluation, ?id=98
- [file] => ?f=99
- [mime] => application/pdf
- [size] => 246992
- [pubname] => 6th European Conference on Computational Chemistry
- [pubinfo] => European Association for Chemical and Molecular Sciences
- [pubkey] => Poster 47
- [workinfo] => September 3-7
- [year] => 2006
- [title] => Online system for molecular descriptors family on structure-activity relationships: assessment and characterization of biologic active compounds
- [authors] => Lorentz JÄNTSCHI, Sorana D. BOLBOACĂ
- [abstract] =>
Starting with an original methodology of molecular descriptor family (MDF) on structure-activity relationships , which was applied to sets of biological active compounds, an open system was created and assessed in order to provide a virtual experimental environment useful in characterization of compounds activities.
A number of thirty-one samples of biologic active compounds were studied by the use of MDF methodology. The best performing models (models with highest abilities in estimation and prediction) were integrated into an online system.
The system integrates the models of thirty one sets of biological active compounds. The range of sets sample sizes vary from 8 to 209 with an average of 45.71 (95% CI [24.87, 66.55]). The previous reported models had an average of the squared correlation coefficients of 0.86 (95% CI [0.82, 0.90]), and a median of the number of variables used equal with 4. The MDF models had an average of squared correlation coefficients equal with 0.90 (95% CI [0.87, 0.92]), an average of cross-validation leave-one-out scores equal with 0.88 (95% CI [0.85, 0.90]), and a median of variable equal with 2. The best performing MDF models had significantly greater correlation coefficients comparing with the previous reported models (p < 0.05).
The open system provide effective models which can be used in studying the activity of new compounds in real time, without any experiments, and with low costs, being necessary just building up as *.hin files the three dimensional structure of the new compound. The future development of the system will allow the access to exhaustive sets of compounds, opening a new pathway in study of theirs activities.
. Jäntschi L. Leonardo Electronic Journal of Practices and Technologies 6 (2005) 76-98.
- [keywords] => Structure-activity relationships; Molecular Descriptors Family; Open system
- [acknowledgment] => Thanks to Ivan CERNUSAK for his pertinent reviews of the full paper. It helps a lot!