You are in: Home > Partners > PP1 - KI

PP1 - KI - National Institute of Chemistry - Laboratory of Chemometrics


National Institute of Chemistry (NIC) is leading research institution in the field of chemistry and related disciplines in Slovenia. It is located in the Slovenian capital Ljubljana and was established in 1946. Today has over 280 employees. The young researchers represent about 25% of the NIC staff members, making NIC one of the leading Slovenian organizations for graduate-level education and training. Many postdoctoral and senior researchers are also involved in academic lectures on the Slovenian Universities. Performed ISO 9001 standards into research and managing of the modern infrastructure and working protocols enables top-level scientific achievements and transferring knowledge into practical use. Basic and applied research are oriented towards life science and material engineering fields: biotechnology, environmental protection, structural and theoretical chemistry, analytical chemistry, materials research, and chemical engineering, through which the institute is in line with the needs of the domestic/international chemical, pharmaceutical, tire, and food industries. In general NIC has 7 laboratories/departments regarding life science research and 9 laboratories regarding chemical analysis and material engineering (
Laboratory of Chemometrics was established in 1983 and is one of 16 laboratories acting on NIC. Currently it has 16 researchers. Our main task is the development chemometrical tools and its promotion to users from academia, industry and regulatory bodies. Regarding partners needs we are developing the tools for modelling of analytical data applicable also in other areas like, optimization of industrial processes, food traceability, analysis of water quality, and modelling of biochemical properties of molecules.


Laboratory of Chemometrics has various range of expertise:
The performance of on demand problem-specialized analysis by using chemometrical - statistical and mathematical tools (e.g. Quantitative Structure Property/Activity Relationships - QSPR/QSAR models). We have long experience in application of advanced modelling methods, to build either deterministic or heuristic models. We have the software and algorithms for different artificial neural networks, for modelling of 3D representations of chemical structures, for calculation of molecular descriptors, for statistical analysis (regression modelling, SVM, PCA).
The Laboratory owns prototypes of validated QSPR/QSAR models for predictions/ classifications of different endpoints of interest in environmental studies (aquatic toxicity of phenols, endocrine disruption activity, water traceability) or health/drug design studies (carcinogenicity of compounds from DSSTox database, toxicity of conazoles, antituberculosis activity of fluoroquinolones, bilitranslocase transport activity, molecular reactivity, antiprion activity). Novel predictions of activities/properties of compounds of interest are available upon request to authors. Input data needed for prediction depends on model, but for most models 2D chemical structures are needed. In case that object of the model is not a chemical structure, other multivariate numerical data are needed.
We have novel prototype for the prediction of α and β transmembrane regions. The innovative part of our algorithm is the possibility of predicting transmembrane (TM) regions of proteins based only on amino acid sequence information that is encoded using mathematical descriptors. The unique feature of our algorithm is that it is independent of evolutionary information and physiochemical property indices, which offers more flexibility than the methods based on evolutionary information. The single sequence based approach of our prediction algorithm is therefore advantageous in predicting for novel non-homologous protein sequences. The prediction algorithms developed using Perl scripts are presently only available upon request to authors.
We have one patent granted and one patent application submitted in collaboration with Slovenian road construction industry regarding new technology of adding the PMMA/ATH dust (waste material from production of composite materials) to bitumen or asphalt mixture.


The basic Laboratory’s activity is application of chemometrics methods in various research fields. In the last period the focus of research has moved from the application of chemometric methods in the analytical chemistry for quality control and quality assurance towards the exploration of large multivariate data sets including the applications in life sciences, such as the evaluation of biochemical and toxicological parameters, which are necessary for chemical regulation (risk assessment), and drug development, accompanied by the studies of adverse effects of drugs/new chemical compounds. We develop 3D representations of chemical structures on the base of aminoacid sequence and NMR data (case study: transmembrane protein bilitranslocase). Using the neural networks, we can develop Quantitative Structure Activity Relationships (QSAR) models for classification and prediction of property/activity for large set of chemical compounds. In regulatory chemistry we use QSAR models, which follow all OECD principles, for prediction of physico-chemical, toxicological and eco-toxicological parameters, which are necessary for risk assessment of chemicals. We are involved in implementation of QSAR methods for European Union Regulation REACH (Registration, Evaluation, Authorisation and Restriction of Chemicals) as an alternative method to animal experiments. In rational drug design we applied QSAR models to search for new leads and to optimize the active substances.


neural network, chemical descriptors, QSAR, statistical analysis, transmembrane protein structure


EU founded projects: TEMPUS, COPERNICUS, IMAGETOX (Intelligent Modelling Algorithms for General Evaluation of TOXicities, FP5), TRACE (Tracing Food Commodities in Europe, FP6), IBAAC (An Integrated Biomimetic Approach to Asymmetric Catalysis), VICIM (Virtual Institute of Chemometrics and Industrial Metrology, FP5), COST D2 (New fluorous media and processes for cleaner and safer chemistry), CAESAR (Computer-Assisted Evaluation of industrial chemical Substances According to Regulations, FP6), COSMOS (Integrated In Silico Models for the Prediction of Human Repeated Dose Toxicity of Cosmetics to Optimise Safety, FP7), BioChemLig (Bio-Orthogonal Chemo-Specific Ligation, FP7), Trans2Care (Transregional Network for Innovation and Technology Transfer to Improve Health Care, European Regional Development Found).

National Institute of Chemistry - Laboratory of Chemometrics
  Marjana Novic
Hajdrihova 19 POB 660 1001 LJUBLJANA Slovenia