Search

A library of thrombin inhibitors has been designed using statistical molecular design. An aromatic scaffold was used, with three varied positions corresponding to three pockets at the active site of thrombin (the S-, P-, and D-pockets). The selection was performed in the building block space, and previously acquired data were included in the design procedure. The design resulted in six, four, and six building blocks for the first (S), second (P), and third (D) pockets, respectively. A second round of selection applied to the combined selected building blocks resulted in a subset of 18 compounds. The selected library was synthesized in parallel and biologically evaluated. The compounds were analyzed with respect to their inhibition (pIC50) of thrombin; membrane permeability, estimated by migration behavior in micellar media (CE log k‘) and pKa; and specificity with respect to inhibition (Ki) of trypsin. Multivariate QSAR studies of the responses yielded valuable results and information that could only be found using statistical molecular design in combination with multivariate analysis.

The reduction of the size of a combinatorial library can be made in two ways, either base the selection on the building blocks (BB's) or base it on the full set of virtually constructed products. In this paper we have investigated the effects of applying statistical designs to BB sets compared to selections based on the final products. The two sets of BB's and the virtually constructed library were described by structural parameters, and the correlation between the two characterizations was investigated. Three different selection approaches were used both for the BB sets and for the products. In the first two the selection algorithms were applied directly to the data sets (D-optimal design and space-filling design), while for the third a cluster analysis preceded the selection (cluster-based design). The selections were compared using visual inspection, the Tanimoto coefficient, the Euclidean distance, the condition number, and the determinant of the resulting data matrix. No difference in efficiency was found between selections made in the BB space and in the product space. However, it is of critical importance to investigate the BB space carefully and to select an appropriate number of BB's to result in an adequate diversity. An example from the pharmaceutical industry is then presented, where selection via BB's was made using a cluster-based design.

With few exceptions, the signal peptides from proteins inserted into, or translocated through, the membranes of gram‐negative bacteria or the endoplasmic reticulum of eukaryotes have no sequence homologies. Therefore these signal peptides have not been considered to contain information related to the different final localizations of the proteins. In this study, 43 signal peptide amino acid sequences from proteins with different final localizations in Escherichia coli have been subjected to a multivariate data analysis. Each amino acid residue was characterized by 20 physico‐chemical properties, yielding a multivariate property profile for each peptide. The similarities/dissimilarities in the property profiles for the signal peptides from different classes were compared with each other by generating few‐dimensional partial least squares (PLS) discriminant plots. With this approach, signal peptides from proteins localized to the periplasmic space (PS), the outer membrane (OM), and the extracellular surroundings (excreted proteins), were separated into distinct groups. Signal peptides from pili proteins were not separated from the OM signal peptides and only partly from the PS signal peptides, but were clearly different from the signal peptides of the excreted proteins. Signal peptides from inner membrane proteins were similar to those of the PS peptides. The size and the hydrophobicity of different peptide segments were responsible for the separation of the signal peptide classes. For example, the hydrophobicity of the N‐terminal segment of the signal peptides increased with an increased distance from the cytoplasm of the final localization for the corresponding proteins. Thus, many signal peptides from proteins with different final localizations in E. coli have different discernible physico‐chemical profiles.

In this study 87 amino acids (AA.s) have been characterized by 26 physicochemical descriptor variables. These descriptor variables include experimentally determined retention values in seven thin-layer chromatography (TLC) systems, three nuclear magnetic resonance (NMR) shift variables, and 16 calculated variables, namely six semiempirical molecular orbital indices, total, polar, and nonpolar surface area, van der Waals volume of the side chain, log P, molecular weight, and four indicator variables describing hydrogen bond donor and acceptor properties, and side chain charge. In the present study, the data from a previous characterization of 55 AA.s from our laboratory have been extended with data for 32 additional AA.s and 14 new descriptor variables. The new 32 AA.s were selected to represent both intermediate and more extreme physicochemical properties, compared to the 20 coded AA.s. The new extended and updated principal property scales, the z-scales, were calculated and aligned to previously reported z(old)-scales. The appropriateness of the extended z-scales were validated by the use in quantitative sequence−activity modeling (QSAM) of 89 elastase substrate analogues and in a QSAM of 29 neurotensin analogues.

The determination of drug‐protein binding parameters (n’s and K’s) can lead to important information on the required therapeutic dosage regimen and possible clinical complications associated with competitive displacement of one drug by a concurrently administered agent. Graphical and computer estimates of the data are often incorrectly formulated, and seldom are adequate data obtained at low binding ratios. Commonly used graphical procedures, inadequately formulated computer methods, and a statistically correct computer method were used to compare results obtained from a circular dichroic examination of dicumarol–human serum albumin and fenoprofen–human serum albumin interactions. Literature binding constants for dicumarol–albumin range from 1 × 105to 30 times that figure, and it is shown here that a wide range in parameter estimates may be obtained depending on the method of data analysis. The parameter estimates in the case of fenoprofen–albumin are even more variable.

Some analytical functions were tested for goodness of fit to standard curves in radio-immunoassays of insulin (Hales & Randle, 1963). Letting a, b and c be parameters, and y the radioactivity measured at an insulin concentration of x, excellent fit was obtained with the function 1/(y+c)=ax+b. This function was associated with a smaller variance around y and a greater number of sign changes for the residuals (calculated radioactivity minus observed radioactivity) than any of the following: 1/y=axc+b; y=aebx+c; y=a+bx+cx2+dx3. The mean number of sign changes of the residuals in 50 separate assays of mouse insulin was within the 95% confidence interval expected for a correct model. The function that was found to be applicable to assays of mouse, human and ox insulins should be useful in routines for computer processing of insulin assay data and may also prove applicable to other radio-immunoassays.