Competencies

Data mining
We have more than ten years experience with analyzing data to identify patterns or relationships. The actual method or methods used depend on the structure of the data but usually lie within the areas: Machine learning (e.g. neural networks, decision trees, genetic algorithms, support vector machines); dimensionality reduction (e.g. clustering, principal component analysis, independent component analyses); and hypothesis testing (e.g. parametric tests like t-test and ANOVA, and non-parametric tests like Wilcoxon rank-sum test and Kruskal-Wallis test).

Biology / bioinformatics

High-throughput technologies like sequencing, microarrays, two-hybrid assays, etc. have created large amounts of data which can only be effectively analyzed using data mining methods as mentioned above. We have extensive experience with mining data generated with such technologies as well as integrating the results with knowledge from a range of biological databases (sequences, annotations, interactions, pathways, publications, etc.).

Computer science

Depending on the task, we develop applications using statistical programming languages like SAS and R, dynamic programming languages like Python, Perl, or programming languages optimized for speed like Fortran and C. We can thus streamline and optimize existing data analysis methods for any situation as well as develop new methods.

Example: Microarray analysis

Many projects requires the application of all of these competencies. When analysizing data from microarrays, for example, the identification of groups of "interesting genes" (data mining) usually needs to be combined with knowledge about the function of these genes and how they interact with each other (biology). Also, microarray analysis is a very active research area with new tools being developed all the time. We monitor these developments and implement new tools regularly (computer science).