Application of DryLab G/plus to the separation of the 30S ribosomal proteins.

If the dependence of retention on temperature is specified for the various components of a sample in isothermal gas chromatography (GC), it is possible to predict retention, bandwidth, and resolution for programmed-temperature GC separations as a function of experimental conditions. The use of a linear-elution-strength (LES) approximation for isothermal retention allows these predictions to be carried out more easily and conveniently, in turn facilitating rapid simulations with a personal computer. This approach to GC method development appears promising, especially if segmented-temperature programs are used. The LES approximation also provides added insight into how different factors affect separation in programmed-temperature GC.

A computer program (DryLab MP®) is described for the simulation of HPLC separations where two or more variables (e.g., temperature and pH) are changed simultaneously. It is assumed that preliminary method development has resulted in a mobile phase of appropriate strength, such that 1 < k′ < 20 for all bands in the chromatogram. If it is desired to simulate separation as a function of changes in the values of n variables, then one or two additional runs are carried out for each variable —changing only that variable. Retention times for each of the latter runs are entered into the computer, and predictions of separation as a function of all conditions are now possible. Because simultaneous changes in two or more variables can lead to significant interaction effects and less accurate predictions, the software evaluates each simulation for possible errors. Allowed conditions must be capable of predicting values of α with an accuracy better than ± 2% (1 S.D.). The present computer program has a number of possible applications during and following method development, as discussed in the following paper (Part II).

A computer program (DryLab® MP) is described that allows restricted multi-parameter mapping for any number (or kind) of separation variables, based on only a few experiments. Multi-parameter computer simulation can be used to develop an high-performance liquid chromatographic method from the beginning, or it can be used to enhance a method developed by other means, e.g., by trial-and-error, single-parameter mapping, etc. The software can also be used to evaluate (and improve) method ruggedness. Finally, various problems (column-to-column variability, change of retention with ambient temperature fluctuations, experimental errors, etc.) that are commonly encountered during routine operation can be handled in the same general way. Examples of these various applications are given.

Computer simulation was used to optimize high-performance liquid chromatography of phenol and its chloro and nitro derivatives. On the basis of two linear gradient runs of different steepness (RP-18—water + methanol + 1% acetic acid), several simulated gradient runs allowed the optimum gradient programme and flow-rate to be chosen so that the time of analysis could be considerably shortened. Good agreement between simulated and experimental chromatograms was obtained in spite of changes in experimental conditions.

Brief review of currently available expert system and simulation software approaches.