Jump to content
Powered by

GENEDATA SCREENER® – specialised data analysis for the identification of high-potential lead structures

Genedata AG has developed a modular software system known as Genedata Screener® that is tailored to screening and hit-to-lead applications. It also enables the automated analysis of physical high-throughput substance screens and the scientific evaluation of the results. Dr. Timo Wittenberg, who works at the company’s site in Constance, Germany, told us about the advantages of the software for identifying high-potential leads.

Dr. Timo Wittenberger studied biochemistry in Tübingen and Tuscon (USA) before obtaining his PhD from the Centre of Molecular Neurobiology in Hamburg (ZMNH). He has worked with Byk-Gulden/Altana Pharma (today Nycomed) since 2001, first in the Bioinformatics division and subsequently in the Discovery-IT department. He has been part of Genedata AG’s “Professional Services” team since 2007. The company, which specialises in the development of software solutions and services, is headquartered in Basel, Switzerland and supports life science companies in the development and use of highly specialised, data-intensive technologies.

Dr. Wittenberger, what are the advantages of the Genedata Screener® platform in the identification of leads?

The Screener can be used for the entire screening process, from validation, structuring and interpretation of data obtained in primary screening processes to the search for highly effective lead structures in the hit-to-lead process. The Genedata Screener integrates the tasks that previously had to be done manually for every single substance, into an automated and standardised workflow.
Genedata Screener® identifies, amongst other things, faulty microtitre plates. © Genedata
The software particularly facilitates the analysis of data from complex screens. It comprises data capture and processing, quality control and assurance and it identifies errors and determines the activity and strength of substances used in different biological tests. The Screener can use data from low-throughput screens as well as data from ultra high-throughput screens. It is also compatible with complex high-content screening technologies.

What is the raw material/data used by the Screener and is there any connection with physical screening?

The Screener uses data acquired by plate readers, i.e. the physical measurement stations installed at the screening sites. The Screener works independently of the technological equipment of the robot systems. The chemical structures originate from the respective company databases and substance libraries.

Can you describe to me in more detail how the software works?

The Screener system consists of four different modules, Assay Analyzer, Condoseo, Kinetics Analyzer and Hit Profiler, each of which have different tasks, analyse different types of data and thus provide information on the most effective lead structures. The modules are based on different algorithms and statistical methods, they employ visualisation tools and the choice of module to be used depends on the requirements of each project and its problem definitions.

The Assay Analyzer, for example, processes raw data and ensures data quality, i.e. it captures data from diverse microtitre plates, identifies artefacts and errors that have occurred during experiments (for example, pipetting errors or other typical problems). Identification is based on statistically significant divergence from the measured data, inconsistencies and alterations in the data pattern that cannot be explained by experimental design. An important advantage over other software tools is that, early in the discovery process, the Screener identifies a large number of false-positive hits and false-negative measurements and thus contributes to time- and cost-savings.

What results are obtained from the Condoseo and Kinetics Analyzer modules?

Condoseo is particularly good at very efficiently recording, calculating and presenting dose-response curves. The system makes it possible to display the inhibition values for different substances tested independently of their concentration. Thus, researchers can obtain ranked lists of the most potent compounds with the best IC50 values, i.e. the concentration in which the target is 50 per cent inhibited. Condoseo also gives the results of secondary and counter screens (screens with similar receptors), thus providing information about chemical structures in the form of numerical values, curves and regression analyses. The Kinetics Analyzer is used for the kinetic analysis of the assays and enables the processing, visualisation and quality control of kinetic assay data. For example, it measures the decrease in fluorescence. The measurement results are converted into diagrams and curves, with the same aim of identifying false-positive and false-negative hits.
Six plots of screening data analysis made for three molecules
Biochemical data and pharmacological parameters are analysed by the HIT Profiler module of the Genedata Screener. © Genedata

Does the system also capture additional parameters, for example, the absorption, distribution or elimination of substances, which are important for the in vivo efficacy of the drug?

This is done by the integrated Hit Profiler. This software module provides a compilation of relevant biological and chemical data from heterogeneous data sources as well as pharmacological parameters of the tested substances, including information on the chemical structure and their variation, which is displayed in two-dimensional images and response curves, as well as information about toxicology, absorption and metabolising enzymes. The Hit Profiler enables the selection of the most effective substances and is equipped with efficient filters for systematic hit and lead identification according to different criteria and properties such as solubility or key structural properties. The system enables the evaluation of tens of thousands of potential hits within a few minutes and also supports the simultaneous analysis of data obtained in previous screens.

Can you give me some more information about the statistics and algorithms used by the Genedata Screener?

The Screener software actually uses the entire range of statistical methods available, including robust methods that help produce reliable results despite the inevitable measurement errors that occur in such large-scale experiments. For example, replicate measurements do not use the mean value as the final result, instead they use the median, which is more likely to diminish individual measurement errors. The processing of dose-response curves involves the weighting of data points according to specific methods, which means that individual measurement errors do not have a major effect on the overall result.

How can the software create cost savings in screening projects?

Our clients have shown in several comparative studies that the use of the Screener delivers superior results (i.e. fewer false-positives and false-negatives). This saves a lot of money because subsequent experimental costs can be drastically reduced. The costs are also reduced because a module such as the Assay Analyzer is an “on-the-fly” solution, which means that the data are analysed during the actual physical screening process. If the Analyzer detects a fault on a microtitre plate, the researcher can immediately stop the process. The ability to process thousands of plates and hence millions of data points per experiment in a single step and to process thousands of dose-response curves in just a few minutes, also saves a great deal of time in the computer-assisted analysis of screening processes, in particular in terms of staffing costs. In addition, standardised processes also contribute to secondary cost savings.

Are there assays or targets that the Screener cannot process or test?

The system is compatible with all kinds of targets (e.g., G-protein-coupled receptors, kinases or proteases) as well as all in-vitro assays involving microtitre plates.

How many screens can be processed with the Genedata system and how long do they take?

There is no limitation in terms of the number of screens. In general, the length of an analysis always depends on the size of the substance library; some of our clients regularly test substance libraries involving several million substances. Calculation capacity is not a problem; however, experimental execution might be a limiting factor. The establishment of an assay can take up to several months; in contrast, the actual screening process, involving modern robotics, can usually be done within a few days. The entire data analysis, including calculation and dose-response curves and lead candidates can be done in a few hours.

How reliable is the screening method and how high is the error rate?

Of course, the final result depends very much on the quality of the measurement data and the reliability of the assays. Therefore, I prefer not to talk about a computer-based error rate, but rather about experimental error rates. The software guarantees optimal analysis of the data. A number of comparative studies of our clients have shown that the quality of the results improves considerably thanks to this software (e.g., halving of the false-positive and false-negative rates).

What are Genedata’s goals in terms of further development of the Screener?

Our major objective is the broader support of new technologies such as high-content screening, i.e. the processing of large quantities of additional information that lead to terabytes of data. Since HCS technology is based on fluorescence measurements and images are recorded, future software systems must be able to recognise the borders of cells or their migration on the basis of the images available. All of us at the Genedata headquarters in Basel are currently working hard to further develop these technologies.

mst – 25 Sept. 2008
© BIOPRO Baden-Württemberg GmbH
Website address: https://www.gesundheitsindustrie-bw.de/en/article/press-release/genedata-screener-specialised-data-analysis-for-the-identification-of-high-potential-lead-structures