Showing papers by "Perry D. Haaland published in 2000"
Patent•
26 Jun 2000TL;DR: In this paper, a method for identifying a component of culture medium based on the parameters (e.g., physical, chemical, biological and/or topological characteristics) of compounds from within a compound library is presented.
Abstract: Methods, apparatus and computer program products are provided for identifying a component of culture medium based on the parameters (e.g., physical, chemical, biological and/or topological characteristics) of compounds from within a compound library. In preferred embodiments, the compound library is a peptide library. Also provided are methods of selecting a compound library from a larger compound space based on whole molecule (i.e., global) parameters of the compounds. Preferably, this method is practiced in conjunction with a method of identifying a component of a culture medium. Further provided are methods of predicting a biological activity of a peptide based on at least one whole molecule parameter of the peptide. This method finds use in methods of drug discovery, identifying components of culture medium, and identifying and/or designing peptides with particular pharmacological or therapeutic activities.
2 citations
Patent•
08 Sep 2000
TL;DR: In this paper, an approach and methods for analyzing whole blood leukocyte response data are disclosed. But, they do not provide any analysis techniques to analyze the data in the database.
Abstract: Apparatus and methods are disclosed for analyzing whole blood leukocyte response data. In a first stage, leukocyte response data are included in a database. In a second stage, various analysis techniques are provided to analyze data in the database.
2 citations
01 Jan 2000
TL;DR: This paper explores a technique for variable and model selection using R-splines, a recently proposed extension to thin plate splines with a modiication to the roughness penalty that allows for a reduced polynomial component to be t.
Abstract: Motivated by an application with a moderate number of potential explanatory variables, this paper explores a technique for variable and model selection using R-splines. R-splines are a recently proposed extension to thin plate splines with a modiication to the roughness penalty that allows for a reduced polynomial component to be t. The key model selection idea is a two-stage approach. First, the important explanatory variables are identiied using a speciic type of R-spline. Then these variables are used to t diierent R-spline models from which the most desirable is chosen. This new method is then compared to all subset regressions by leaps and bounds and regression trees. An application of the methodology is also discussed.