Latest News


Systems Medicine Conference in Slovenia

Systems Medicine Conference in Slovenia: National Awareness Event,12th CFGBC Symposium and “Systems Medicine” Workshop took place from June 8th-9th 2017 at Hotel Slon, Ljubljana, Slovenia.

CASyM winter school of Systems Medicine

CASyM winter school of Systems Medicine took place between March 29th and April 1st 2017 in Ljubljana, Slovenia and is entitled »The 3rd SysBioMed hands-on tutorial: Systems Medicine Approaches in Personalized Medicine«

ELIXIR-SI launch & 11TH CFGBC Symposium

The event with motto “Data for Life” took place from September 20th - 21st 2016 at Cankarjev dom and the Faculty of Medicine, University of Ljubljana, Slovenia.

Data4Life 2016: Abstracts of Posters

Abstracts of Posters

Identification of molecular and metabolic differences between NASH and HCC of Cyp51 knockout mice

Kaja Blagotinšek1, Gregor Lorbek1, Žiga Urlep1, Martina Perše2, Jera Jeruc3, Peter Juvan1, Damjana Rozman1

1Centre for Functional Genomics and Bio-Chips, Institute of Biochemistry, Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia, 2Medical Experimental Centre, Institute of Pathology, Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia, 3Institute of Pathology, Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia

Introduction: Nonalcoholic fatty liver disease (NAFLD) encompasses numerous histopathological features, ranging from simple steatosis to steatohepatitis (NASH), which is characterized by a more serious form of liver damage and fibrosis. NASH can progress to inflammation- and oxidative stress-induced cirrhosis and even hepatocellular carcinoma (HCC) [1,2]. Based on the literature, up to 29% of HCC cases could result from NAFLD [3]. The purpose of this study is to identify the differences between NASH and HCC with regard to gene expression and metabolism that may provide deeper insight into the progression of disease.

Methods: Experiments were performed on the hepatocyte knockout mouse model of lanosterol 14α-demethylase (HCyp51-/-) to evaluate the age-related changes due to cholesterol imbalance in the liver. Since liver is a sexually dimorphic organ we performed experiments on animals from both sexes aged 1, 1.5 and 2 years. Special focus was paid to molecular mechanisms which are important for the initiation and progression from NASH to HCC. Histopathology was performed on H&E liver sections. Biochemical parameters were measured in plasma. Gene expression profiling was conducted by qPCR. The data was processed and analysed using two-way ANOVA test.

Results: At the ages of 1 year and older, the HCyp51-/- mice exhibit advanced chronic inflammation and fibrosis, which are characteristic for NASH. In our model NASH progresses to HCC with an age-dependent liver pathology and sexual dimorphism, with a HCC female to male ratio of 2 to 1. Liver is a sexually dimorphic organ with crucial metabolic pathway differing between females and males. Special focus is paid to hepatocyte cholesterol synthesis in the liver, where fundamentally up-regulated cholesterogenic genes (Sqle, Lss, Nsdhl, Tm7sf2) in 1.5 year old males were observed. Lipid plasma parameters were mostly unchanged, however, elevated ALT and AST ratio in KO mice pointed to liver injury.

Conclusions: Mechanisms involved in HCC development in HCyp51-/- mice include a combination of chronic complementary effects: the block of the hepatocyte Cyp51 reaction resulted in deregulation of the sterol network, which promotes inflammation and ER stress. At the same time it leads to metabolic deregulation due to deficiency in post-lanosterol RORC ligands. Our future perspective is to assess how this chronic inflammatory environment with leukocyte infiltration and activated cytokine signalling influences on RORC, WNT and HGF signalling and find major hallmarks in transition from NASH to HCC.

Acknowledgements: This work was supported by the Slovenian Research Agency (ARRS) program grant P1-0390. K. Blagotinsek is supported by the graduate fellowship of ARRS.


1. Machado VM and Diehl MA (2016) Gastroenterology 150 (8), pp 1769-77.

2. Ma C et al. (2016) Nature 531, pp 253-7.

3. Clark DJ, et al. (2014) Dig Dis Sci 59(2), pp 365-74.

Use of Steatonet as a predictive tool to disclose the complexity of gender-based liver-related diseases

Tanja Cvitanović1, Miha Moškon2, Miha Mraz2, Damjana Rozman1

1Centre for Functional Genomics and Bio-chips, Faculty of Medicine, University of Ljubljana, Slovenia, 2Faculty of Computer and Information Science, University of Ljubljana, Slovenia

Introduction: Almost 200 reactions are used by computational model SteatoNet to describe complex interactions of the liver with peripheral tissues. Metabolic, gene regulatory and signal transduction pathways are all included within the model. Due to its object-oriented nature, it can be easily adapted to liver-associated pathologies such as non-alcoholic fatty liver disease. Furthermore, it can be used to perform dynamical simulations even in the case of sparse experimental data (1).

Problem: Liver is one of the most sexually dimorphic organs, surpassed only by testes and ovaries. Gender-based differences were discovered also in the livers of mice with hepatocyte deletion of Cyp51, which have and shown how disrupted hepatic cholesterol synthesis reflects differently on the whole body homeostasis in males and females (2).

Results: Current modifications of SteatoNet include simulations of sex hormone levels in the blood and their networking with liver and peripheral tissues. To our knowledge SteatoNet (F/M) represents the first gender-based liver metabolic model. To set one’s sight on the predictive nature of SteatoNet, consequences of the disrupted hepatic cholesterol synthesis in adipose tissue will give an excellent starting point for further experimental testing.

Conclusion: SteatoNet has the potential to be developed further into a diagnostic, predictive or analytical tool geared towards personalized treatment of patients. It additionally shows large potentialcan potentially also to predict the network effects of polymorphisms associated with liver-related pathologies.


1. Naik, A., D. Rozman, and A. Belic, SteatoNet: the first integrated human metabolic model with multi-layered regulation to investigate liver-associated pathologies. PLoS Comput Biol, 2014. 10(12): p. e1003993.

2. Lorbek, G., et al., Lessons from hepatocyte-specific cyp51 knockout mice: impaired cholesterol synthesis leads to oval cell-driven liver injury. Sci Rep, 2015. 5: p. 8777.

Discovering biomarkers of Alzheimer disease

Jure Fabjan1,3, Ursula Prosenc Zmrzljak1, Alja Videtič Praska2, Zvezdan Pirtošek3, Damjana Rozman1

1Centre for Functional Genomics and Bio-Chips, Institute of Biochemistry, Faculty of Medicine, University of Ljubljana, Slovenia, 2Medical Centre for Molecular Biology, Institute of Biochemistry, Faculty of Medicine, University of Ljubljana, Slovenia, 3Neurology Clinic, Ljubljana University Medical Centre, Slovenia

It is estimated that there were 46.8 million cases of dementia in the world in 2015 [1]. The most frequent cause of dementia is Alzheimer disease (AD), which is associated with extracellular amyloid plaques and intracellular neurofibrillary tangles [2]. 95% of cases with AD can be classified as late onset AD (LOAD), of which the biggest genetic risk factor is Apolipoprotein E (ApoE) E4 allele. The carrier of this allele does not necessarily get LOAD, but E4/- heterozygotes have two to three-fold increased risk of LOAD, while E4/E4 homozygotes have five-fold increased risk of LOAD. Carrying this allele is also associated with earlier onset of the disease [2], [3]. The frequency of ApoE E4 allele varies by region and its estimate is the highest in Northern Europe and the lowest in Asia and Southern Europe [3]. The aims of our research are (i) to assess the frequency of ApoE alleles in Slovene AD and non-AD population, as such estimate was not done yet, and (ii) to enable more accurate assessment of risk of AD in Slovenia with incorporation of genetic biomarkers into diagnosis protocol. Test subjects will be recruited in clinic of Centre for neurodegenerative diseases at Neurological clinic from patients with diagnosed AD by established clinical criteria. Controls will be recruited from patients relatives and companions without diagnosed AD and with comparable demographics (sex and age), as the test group. We estimate to have at least 200 subjects in each group. The peripheral blood will be taken from subjects and from it the DNA will be isolated and used for ApoE genotyping. Our project started with application for ethical approval from the Commission of the Republic of Slovenia for medical ethics. At the same time we assembled a database which could be used routinely in diagnostic laboratories for verification of genetic biomarkers. By doing this we expect to achieve traceability of the patients’ tissue samples and their anonymity. After verifying and improving both the database and the protocol, the ApoE genotyping will be performed.


[1] M. Prince, A. Wimo, M. Guerchet, A. Gemma-Claire, Y.-T. Wu, and M. Prina, “World Alzheimer Report 2015: The Global Impact of Dementia - An analysis of prevalence, incidence, cost and trends,” Alzheimer’s Dis. Int., p. 84, 2015.

[2] C. Reitz and R. Mayeux, “Alzheimer disease: Epidemiology, diagnostic criteria, risk factors and biomarkers,” Biochem. Pharmacol., vol. 88, no. 4, pp. 640-651, 2014.

[3] A. Ward, S. Crean, C. J. Mercaldi, J. M. Collins, D. Boyd, M. N. Cook, and H. M. Arrighi, “Prevalence of Apolipoprotein E4 genotype and homozygotes (APOE e4/4) among patients diagnosed with alzheimer’s disease: A systematic review and meta-analysis,” Neuroepidemiology, vol. 38, no. 1, pp. 1-17, 2012.

Facilitating treatment selection in malignant mesothelioma using clinical-pharmacogenetic models

Katja Goričar1, Viljem Kovač2, Vita Dolžan1

1Pharmacogenetics Laboratory, Institute of Biochemistry, Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia, 2Institute of Oncology Ljubljana, Ljubljana, Slovenia

Introduction: Malignant mesothelioma (MM) is an aggressive tumor with poor prognosis. Large interindividual differences in response to standard gemcitabine/cisplatin (GEM/CIS) or pemetrexed/cisplatin (PMX/CIS) chemotherapy present an important issue in MM treatment. Because both clinical characteristics and genetic variability may affect treatment outcome, our aim was to construct and validate clinical-pharmacogenetic prediction models of treatment outcome in MM for both chemotherapy combinations and to develop an algorithm for genotype-based treatment recommendations.

Methods: Clinical-pharmacogenetic models were built on 71 GEM/CIS-treated and 57 PMX/CIS-treated MM patients. Pharmacogenetic scores were assigned by rounding the regression coefficients. GEM/CIS model was validated on 66 independent MM patients.

Results: The model for predicting response to GEM/CIS chemotherapy included CRP level, histological type, performance status, RRM1 rs1042927, ERCC2 rs13181, ERCC1 rs3212986, and XRCC1 rs25487 with scores ranging between 0 and 3.4. Cutoff value of 0.75 had sensitivity of 0.62 and specificity of 0.81. Patients with higher score had significantly shorter progression-free and overall survival (P<0.001). In the validation group, positive predictive value was 0.74 and negative predictive value was 0.56. The model for predicting response to PMX/CIS chemotherapy included CRP level, MTHFD1 rs2236225, and ABCC2 rs2273697 with scores ranging between 0 and 3.9. Cutoff value of 2.7 had sensitivity of 0.75 and specificity of 0.61. Patients with higher score had significantly lower probability of good response and shorter progression-free survival (P<0.001).

Conclusions: Clinical-pharmacogenetic models could enable stratification of MM patients based on their probability of response to GEM/CIS or PMX/CIS, thus facilitating treatment selection. This approach could therefore improve treatment outcome and also enable translation of pharmacogenetic testing to clinical practice.

Nitrosative stress, telomere length and telomere dynamics in juvenile patients with type 1 diabetes

Tinka Hovnik1, Tine Tesovnik1, Jernej Kovač1, Tadej Battelino2, Katarina Trebušak Podkrajšek1

1Unit of Special Laboratory Diagnostics, Children’s Hospital, University Medical Centre Ljubljana, Slovenia, 2Department of Pediatric Endocrinology, Diabetes and Metabolic Diseases, Children's Hospital, University Medical Centre Ljubljana, Slovenia

Type 1 diabetes (T1D) is a disease caused by autoimmune destruction of pancreatic β-cells. Lack of insulin is associated with the chronic state of hyperglycemia, oxidative stress and inflammation and leads to development of diabetic complications. The aim was to evaluate telomere length (TL) and TL dynamics as biomarkers in juvenile patients with T1D in relation to glycemic control and nitrosative stress.

44 juvenile participants with T1D, 17 with poor (HbA1c>9%) and 27 matched patients with good glycemic control (HbA1c<7.5%) were included in the study. 237 chronological DNA samples were assessed for relative TL telomere dynamics. Additionally, nitrosative stress was measured with ELISA assay using 8-Nitroguanine (8-NO2-Gua) as a biomarker in plasma samples of 10 patients with poor glycemic controls and their matched good glycemic controls.

Our study indicate increased nitrosative stress in patients with poor glycemic control (Fig. A), but TL association with glycemic control is not seen probably due to the short investigated period of time. The linear regression of all TL measurements indicates the impact of T1D duration, even after adjustment to the age of participants (Fig. B).

A larger number of chronological samples over longer period of disease duration would be required to confirm telomere dynamics and nitrosative stress in association with glycemic control.

This work was supported by Slovenian National Research Agency grant J3-6800.

Figure 1: Fig. A: The difference in 8-NO2-Gua concentration between T1D patients with poor (c=256.7 ng/mL) compared to good glycemic control (c= 144.8 ng/mL; p=0.0007). Fig. B: The correlation between rTL and duration of T1D adjusted for age (slope= -0.264, r2= 0.077, p<0.0001)

Association of polymorphisms in Aurka and Polo kinases with gastric cancer risk

Petra Hudler1, Ela Markočič1, Aner Mesić2, Marija Rogar1, Radovan Komel1

1Institute of Biochemistry, MCMB, Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia, 2Department of Biology, Faculty of Science, University of Sarajevo, Sarajevo, Bosnia and Herzegovina

Introduction: Gastric cancer is in decline in most developed countries; however, it still accounts for a notable amount of global mortality and morbidity related to cancer. It is associated with H. pylori infection, EBV virus, and diet, as well as with genetic and epigenetic factors. In addition, it has been established that chromosomal instability (CIN) plays an important role in the development of gastric cancer. It was assumed that CIN could develop due to aberrant segregation of chromosomes during mitosis. Polymorphisms in genes implicated in chromosome segregation could affect their expression and could therefore increase the risk of developing cancer.

Aim: The aim of our research was to define the association between polymorphisms in segregation genes, including rs911160 (gene AURKA), rs2289590 (gene AURKB), rs11084490 (gene AURKC) and rs42873 (gene PLK1), and risk of developing gastric cancer. We also evaluated the association between polymorphisms and clinicopathological characteristics of gastric cancer.

Methods: We performed a retrospective case-control study. Genotyping was performed using real-time polymerase chain reaction. (RT-PCR). The results were statistically evaluated.

Results: Polymorphism rs11084490 in gene AURKC was associated with invasion of cancerous cells in lymphatic tissue (F=13,550; p=0,001). This association was most significant for patients with genotype CC. Genotype GG was associated with diffuse type of gastric cancer (F=5,837; p=0,041) and the poorest survival rates (Mantel-Cox test; χ2 =6,557; p=0,038), whereas genotype CG was found commonly in the group of patients with intestinal type of cancer. We also confirmed a strong association between survival of patients with gastric cancer and Lauren classification and perineural invasion.

Conclusions: We confirmed that alleles of polymorphism rs11084490 in gene AURKC were associated with key histopathological features of gastric adenocarcinomas and survival rate of the patients. Polymorphisms, like rs11084490, could be therefore used as prognostic biomarkers.

Normalization of erythropoietin and estrogen receptors detection in breast cancer cell lines

Mateja Cigoj1, Urška Kotnik1, Sara Košenina1, Radovan Komel1, Nataša Debeljak1

1Institute of Biochemistry, Faculty of Medicine, University of Ljubljana, Slovenia


Hormone dependent breast cancer is characterized by estrogen receptors (ER), enabling treatment with an estrogen receptor antagonists tamoxifen. Signalization of ER with other membrane receptors, including erythropoietin receptor (EPOR), presumably plays a role in resistance to tamoxifen [1, 2]. In aim to characterize the level of ER and EPOR receptors on multiple breast cancer cell lines via Western blot, appropriate normalization and validation procedure is necessary.

The most common normalization controls include housekeeping proteins, such as Glyceraldehyde 3-phosphate dehydrogenase (GAPDH), beta-actin and tubulin. However in cancer cells the expression of many proteins including housekeeping proteins may be altered [3]. Therefore it is important to validate with appropriate normalization control. Furthermore antibody validation is crucial to confirm antibody specific and reduce batch-to- batch variations [4].

The aim of our study is to determine appropriate normalisation process for selected breast cancer cell lines and to test different primary and secondary antibodies to ensure specificity and reproducibility of the results.


Total proteins were isolated from breast cancer cell lines (human MCF-10A, MCF-7, T-47D, MDA-MB 361, MDA-MB 231, Hs578T). EPOR, ERa, ERb, membrane G protein-coupled estrogen receptor (GPER) and beta-common receptor (bcR) were analyzed and quantified by standard Western blot analysis. GAPDH was tested as a normalization body. Different quantities of secondary antibodies were tested.

The expression of EPOR was confirmed in all analyzed cell lines, specificity of other receptors varied upon antibodies. The normalization of results using GAPDH is not optimal, because of uneven GAPDH expression among analyzed cell lines. Normalization to total amount of protein was also applied. Normalization with other antibodies is in progress.


Lack of specificity and reproducibility are common problems among commercial antibodies, therefore antibodies that show multiple nonspecific signals should be carefully validated before use. In our further work we will choose appropriate positive and negative controls in order to determine whether the detected proteins corresponds to our target of interest.

To ensure accurate quantitative data from Western blot analysis we need to establish an appropriate normalization procedure. We found that GAPDH is not a suitable normalization control. To overcome this problem we will analyze the expression of some other housekeeping proteins and also attempt normalization to total amount of protein in a sample using a protein stain such as Coomassie brilliant blue.


[1] Trošt N, et al. Int J Mol Med, 2013, 31.3: 717-725.

[2] Ilkovičova L, (2016) Oncology reports (submitted)

[3] Rubie C, et al. Mol Cell Probes, 2005, 19.2: 101-109.

[4] Bordeaux J, et al. Biotechniques, 2010; 48: 197

Design of synchronous transcription based biological memory

Lidija Magdevska1, Žiga Pušnik1, Miha Mraz1, Nikolaj Zimic1, Miha Moškon1

1Faculty of Computer and Information Science, University of Ljubljana, Ljubljana, Slovenia


Attention toward the implementation of robust and scalable biological memory has been increasing in recent years. Despite significant progress in the field of synthetic biological memory, the implementation of scalable synchronous sequential structures in biological systems has yet to be achieved. This type of structures is essential for the implementation of more complex biological information processing structures, since they provide synchronisation between the logic elements, which results in a robust behaviour of the system.


We present a computational design of an edge-triggered D flip-flop based on transcriptional logic. The edge-triggered flip-flop was implemented with a master-slave configuration of the basic clocked D flip-flop. We applied a genetic algorithm to tune the response of the topology with the calibration of kinetic parameter values. We also confirmed the robustness of the topology using a global sensitivity analysis framework developed specifically for the purpose of dealing with high-dimensional and poorly connected parameter spaces, as is the case in the proposed system. The designed structure was then applied to a robust implementation of a Johnson counter, which can count up to 2n events using a sequence of n flip-flops.


The described memory elements are robust and therefore satisfy the requirements of many applications of synthetic biology. The key advantage of our approach is the use of biologically relevant data in the optimization process. This enables the potential construction of biological parts used in the proposed counter topology and, with that, paves the path for an in vivo implementation of biological storage and synchronisation devices.


The research was partially supported by the scientific-research programme Pervasive Computing (P2-0359) financed by the Slovenian Research Agency in the years from 2009 to 2017 and by the basic research and application project Designed cellular logic (J1-6740) financed by the Slovenian Research Agency in the years from 2014 to 2017. Results presented here are in scope of Ph.D. theses that are being prepared by Lidija Magdevska and Žiga Pušnik.

Time-dependent stochastic global sensitivity analysis of gene regulatory networks with multiple transcription factor binding sites

Mattia Petroni1, Miha Mraz1, Nikolaj Zimic1, Miha Moškon1

1Faculty of Computer and Information Science, University of Ljubljana, Ljubljana, Slovenia


Sensitivity analysis methods are widely applied to the assessment of system’s robustness and can be used to efficiently tune the modelled response with the experimental results. The Morris sensitivity analysis is a screening method that has been extensively used for performing the global sensitivity of chemical reaction network models in systems biology [1].


We present an adaptation of the Morris technique based on a One-Step-At-A-Time (OAT) screening experiment for investigating the time-dependent sensitivity of quantitative parameters in the stochastic models of gene regulatory networks (GRNs) with multiple transcription factor binding sites. Models describing these networks can quickly become intractable, because of the exponential number of binding reactions. We efficiently reduced their numbers with the application of the multiscale stochastic simulation algorithm (mSSA) which is able to perform the simulations in the observed systems in a feasible time. Moreover, we further reduced the analysis time with the coarse grained parallelization of the mSSA introducing the parallel mSSA algorithm (ParMSSA). ParMSSA is used by the screening experiment to speed-up the simulation time of multiple mSSAs required to obtain statistically relevant results. We demonstrate the proposed approach by performing the sensitivity analysis on different GRNs with multiple binding sites.


The proposed adaptation of the Morris sensitivity analysis method allows us to efficiently perform the analysis of the complex GRNs additionally taking into account the stochasticity of their dynamic response. Its parallelisation allows us to confront the computational complexity of the Morris sensitivity analysis. This allows us to more accurately identify the most relevant parameters governing the dynamics of the system. Furthermore, it enables us to perform additional model reduction and simplification, and to accurately tune its response with the experimental results taking into account the stochasticity of the model response even in complex GRNs.


The research was partially supported by the scientific-research programme Pervasive Computing (P2-0359) financed by the Slovenian Research Agency in the years from 2009 to 2017 and by the basic research and application project Designed cellular logic (J1-6740) financed by the Slovenian Research Agency in the years from 2014 to 2017. Results presented here are in scope of Ph.D. thesis that is being prepared by Mattia Petroni.

1. Zi Z. (2011) IET Systems Biology 5, pp.336-46.

Global sensitivity analysis of biological models with high-dimensional and poorly connected parameter spaces

Žiga Pušnik1, Lidija Magdevska1, Miha Mraz1, Nikolaj Zimic1, Miha Moškon1

1Faculty of Computer and Information Science, University of Ljubljana, Ljubljana, Slovenia


Sensitivity analysis methods have been widely applied to systems as well as synthetic biology. They are able to assess the system’s robustness, can guide parameter estimation, experimental design and can serve as model validation tool. Sensitivity analysis methods can be divided in local and global methods. While local methods analyse system sensitivity in the local neighbourhood of the nominal parameter value for single parameter at a time, global methods tend to investigate whole space of possible parameter space applying different sampling techniques. Local sensitivity analysis methods have proven to be inefficient when parameter values are either missing, partially known or exhibits large variation. On the other hand, global sensitivity methods fail when viable solution space is small in comparison to the whole space, which is often the case when dealing with high dimensional models.


We present a computational framework that is able to perform the global sensitivity of biological models having high-dimensional and poorly connected viable parameter regions. The framework can be decomposed into the following steps: (1) generating the solutions that describe viable parameter regions using optimisation meta-heuristics, (2) clustering the solutions based on their connectivity into viable parameter regions, (3) parameter sampling within the clusters, (4) assessing the global sensitivity values using generated samples, and (5) merging the sensitivity values for each parameter from each of the clusters into sensitivity values describing parameter influence on broad system dynamic. Viable solutions are obtained with genetic algorithms. The solutions are then clustered together by k-means method. Orthogonal sampling is then performed on each of the clusters and for every sample Morris sensitivity analysis is applied. Finally average means and standard deviations of partial effects are combined for all clusters.


Described computational framework can be efficiently applied to the sensitivity assessment especially when dealing with poorly connected parameter space. It has already been efficiently applied to the analysis of genetic master-slave D flip-flop, for which several unconnected viable parameters regions were found. We were able to apply the results of the framework to identify the most robust topology and determine the parameter regions for which the flip-flop dynamics were optimal.


The research was partially supported by the scientific-research programme Pervasive Computing (P2-0359) financed by the Slovenian Research Agency in the years from 2009 to 2017 and by the basic research and application project Designed cellular logic (J1-6740) financed by the Slovenian Research Agency in the years from 2014 to 2017. Results presented here are in scope of Ph.D. theses that are being prepared by Lidija Magdevska and Žiga Pušnik.

GoMapMan: Plant specific ontology to ease analyses of high-throughput data

Živa Ramšak1, Špela Baebler1, Björn Usadel2,3, Kristina Gruden1

1Department of Biotechnology and Systems Biology, National Institute of Biology, 1000 Ljubljana, Slovenia, 2Institute for Biology I, RWTH Aachen University, D-52056 Aachen, Germany, 3IBG-2: Plant Sciences, Institute for Bio- and Geosciences, Forschungszentrum Jülich, 52425 Jülich, Germany

Introduction: Understanding the different aspects of plant biology using systems biology tools can contribute to crop plant breeding and development of efficient agricultural practices. However, the approach is hindered by lacking or dispersed crop-specific experimental data, functional annotations and visualization tools.

Results: We have developed GoMapMan (, an open web-accessible resource for gene functional annotations in the plant sciences. It was developed to facilitate improvement, consolidation and visualisation of gene annotations across several plant species. GoMapMan is based on the MapMan ontology, organized in the form of a hierarchical tree of biological concepts, which describe gene functions. Currently, genes of the model species Arabidopsis and six crop species (potato, tomato, rice, tobacco, beet and cacao tree) are included. The main features of GoMapMan are 1) dynamic and interactive gene product annotation through various curation options, 2) consolidation of gene annotations for different plant species through the integration of orthologue group information, 3) traceability of gene ontology changes and annotations, 4) integration of external knowledge about genes from different public resources, and 5) providing gathered information to high-throughput analysis tools via dynamically generated export files.

Conclusions: Using GoMapMan, the knowledge on plant biology can be improved by translating existing knowledge from model to crop species and by easier data interpretation of crop experimental data.

References: Ramšak Ž. et al. (2014) Nucleic Acids Res. 42:D1167-75.

Genetic and clinical determinants of response to treatment in Parkinson's disease: can they help to personalize the treatment?

Sara Redenšek1, Maja Trošt2, Vita Dolžan1

1Pharmacogenetics Laboratory, Institute of Biochemistry, Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia, 2Department of Neurology, University Medical Centre Ljubljana, Ljubljana, Slovenia

Introduction: Parkinson's disease (PD) is a complex progressive neurodegenerative brain disorder with increasing prevalence in aging population. Genetic and clinical factors play an important role in the etiopathogenesis of PD and its response to treatment. As the preparative step for the clinical pharmacogenetic study of treatment response in PD we have performed a review of the current literature to decide which clinical and genetic factors should be included in our analysis.

Methods: We performed a PubMed search on factors related to etiopathogenesis of PD and on factors, especially genetic polymorphisms, related to response to anti-parkinsonian drugs.

Results: Regarding etiopathogenesis of PD, several cellular pathways were found to be compromised due to genetic mutations: protein aggregation, protein and membrane trafficking, lysosomal autophagy, immune system, neurodevelopment, neuron cell differentiation and survival, mitochondrial homeostasis and other processes, which may modify the disease course and response to treatment depending on the damaged pathway [1]. In addition, response to treatment may be influenced by genetic polymorphisms in dopamine, neurotransmitter and drug metabolism and transport, dopamine receptors, signalling pathways, inflammation, antioxidative defense and synaptic transmission. Genetic variability could be the reason for occurrence of adverse drug reactions (dyskinesias, motor fluctuations, hallucinations, sleep attacks) and may also affect the appropriate dosing [2]. Beside genetic determinants also several other factors may influence treatment response and disease course: gender, age, age at onset, symptoms before and at the treatment initiation, family history of brain diseases, history of head injuries, accompanying diseases and treatments, living in rural or urban areas, environmental exposures (pesticide exposure, well water drinking) and patient's lifestyle (coffee and alcohol consumption, smoking, physical activity) [1]. Usually, the disease severity at the start of the treatment turns out to be the most significant factor influencing treatment response.

Conclusion: Based on the acquired data we designed a clinical pharmacogenetic study in which we will evaluate the impact of each mentioned factor on the efficacy of anti-parkinsonian drugs as well as the occurrence and time to occurrence of adverse drug reactions. We will analyze the combined effects of clinical and genetic data using up-to-date statistical approaches to identify biomarkers of different treatment outcomes. If such prediction models could be established they would enable personalized treatment approach of PD.

Acknowledgments: We thank prof. dr. Zvezdan Pirtošek, dr. med. for helpfull discussions.


1. Kalia LV, et al. (2015) Lancet 386(9996), pp. 896-912.

2. Schumacher-Schuh AF, et al. (2014) Pharmacogenomics 15(9), pp. 1253-71.

Microbes call for ELIXIR-SI

Robert Šket1, Zala Prevoršek1, Blaž Stres1,2

1Biotechnical Faculty, University of Ljubljana, Ljubljana, Slovenia, 2Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia

Introduction: There are 10 times more microbial cells present within or on the surface of our body than there are our own cells. Microbial counterpart contains 150 times more genes than our genome and hence interact in complex metabolic relationship with the host.

A number of ongoing projects has focused on a more systematic and structured exploration of microbial communities (metagenomes), their gene transcripts (metatranscriptomes) and resulting metabolites (metabolomes) in various types of mammalian intestinal tracts in relation to (i) human physiology or disease. (e.g. Humans - PlanHab (IJS); PreTerm (Faculty of Medicine); IBD- Children's Hospital); (ii) animal physiology or desease or biotechnological applications -Microbial Eznymes (Biotechnical Faculty)) and (iii) environment - WaterCaves (Biotechnical Faculty)).

The resulting datasets from these projects in the range of 20 GB -1 TB are currently being processed either on smaller dedicated servers locally or HPCCs outside ELIXIR.

Results: To comparatively analyze on a large scale the metagenomes, metatranscriptomes, assemble draft genomes of microorganisms from metagenome sequences, analyze draft genomes within the context of already closed genomes and integrate metadata data within Bayesian Networks we call for a more systematic access to ELIXIR-SI based HPC resources in the range of e.g. 500-1000 CPU, 4-8TB RAM per single job. A more systematically dedicated training tailored to the specificities of the datasets and needs of increasing number of biologists and students analyzing the data would boost the development of the microbial systems biology, from the perspective of human health, biotech and environment, fostering discoveries of additional novel enzymes, antibiotics, directed evolution of proteins or transcripts coded in metagenomic or metatranscriptomic DNA for industrial applications.

Conclusions: We call for a more focused discussion on the needs of this part of scientific research faced with large-scale data, complex metabolic and co-occurrence networks, multifactorial responses and multivariate datasets with the idea to make ELIXIR-SI functionally and operationally more accessible to researchers performing the analyses in microbial 'omics areas.

Reanalysis of bacterial 16S rRNA in rumen of wild animals for microbial pathogen signatures: ELIXIR-SI implementation of MOTHUR and R

Robert Šket1, Jernej Porenta2, Jan Jona Javoršek3, Blaž Stres1,4

1Biotechnical Faculty, University of Ljubljana, Ljubljana, Slovenia, 2ARNES, Ljubljana, Slovenia, 3Institut Jozef Stefan, Ljubljana, Slovenia, 4Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia

Introduction: Ruminants are one of the most successful groups of herbivorous mammals on the planet with around 200 species represented by 75 mio wild and 3.5 bil domesticated animals worldwide. The ruminal microbial assemblages are remarkably diverse, containing hundreds of different bacterial, archaeal, protozoal and fungal species capable of plant residues degradation and conversion into plethora of valuable products. Bacterial microbial communities represent the most diverse and functionally resilient part of communities in rumen due to interaction with highly diverse plant residues in feed.

Results: C++ program MOTHUR (Schloss et al., 2009) and its accompanying bacterial, archaeal, protozoal (SILVA ( and fungal (UNITE ( databases were installed next to the relevant R routines (vegan) for streamlined statistical analyses of the training set composed of 1.5 mio sequences.

Conclusions: The implementation enabled (i) testing the established routines for subsequent high-throughput of analyses of additional datasets derived from deep amplicon sequencing, (ii) provided grounds for testing for the presence of microbial pathogens of medical importance for humans (e.g. animal related disease(s), harbouring multiple antibiotic resistance genes) and (iii) assessing the effects of systematic errors related to the use of different sequencing technologies.

Figure 1: Figure 1: An example of NM-MDS visual representation of factors affecting the structure of bacterial microbial communities obtained from two different sequencing platforms and corrected for non-overlapping stretches of sequenced genes over time.

Deficient RORC activity contributes to metabolic insufficiency in a subset of Cyp51 knockout mice

Žiga Urlep1, Gregor Lorbek1, Martina Perše2, Jera Jeruc3, Peter Juvan1, Madlen Matz-Soja4, Rolf Gebhardt4, Ingemar Björkhem5, Jason A. Hall6, Alexander Rives7, Richard Bonneau7, Dan R. Littman6,8, Damjana Rozman1

1Centre for Functional Genomics and Bio-Chips, Institute of Biochemistry, Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia, 2Medical Experimental Centre, Institute of Pathology, Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia, 3Institute of Pathology, Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia, 4Institute of Biochemistry, Faculty of Medicine, University of Leipzig, Leipzig, Germany, 5Department of Laboratory Medicine, Division of Clinical Chemistry, Karolinska Institute, Karolinska University Hospital, Huddinge, Sweden, 6The Kimmel Center for Biology and Medicine of the Skirball Institute, New York University School of Medicine, New York, New York 10016, USA, 7New York University & Simons Foundation for Data Analysis, New York, NY 10010, USA, 8Howard Hughes Medical Institute, New York University School of Medicine, New York, New York 10016, USA

Background: Initially identified as an orphan nuclear receptor, RORC has in recent years been shown to exhibit pleiotropic effects, ranging from fine-tuning of the circadian clock, immune system development and hepatic metabolic control. It was recently discovered that cholesterol biosynthesis intermediates (CBI) after Cyp51 mediated demethylation serve as natural RORC ligands in vitro. By applying the hepatocyte Cyp51 knockout mice (HCyp51-/-) we aimed to examine whether sterols modulate RORC transcriptional activity also in vivo.

Methods: CBIs were measured by GC/MC. Expression profiles of HCyp51-/- and control mice livers were examined by Affymetrix microarrays. The RORC transcriptional profiles were assessed by RNA sequencing of the hepatocyte Rorc knockout mice (HRorc-/-). Comparative transcriptome analysis was conducted in R, using Bioconductor packages limma and pgsea.

Results: We found decreased post-CYP51 CBIs in 6- and 19-week HCyp51-/-mice, particularly evident in a subgroup of mice with the worst observable phenotypes (runts). Gene set enrichment analysis exhibited decreased RORC transcriptional activity in runts and 19-week HCyp51-/- mice. To improve the selection of RORC targets, transcriptional profiles of HCyp51-/- mice were compared to those of HRorc-/- mice. The circadian nature of Rorc expression permitted the identification of the proposed Rorc target genes, approximately 30% of which were deregulated also in HCyp51-/- mice. These genes are involved mainly in amino acid metabolism as shown by GO and KEGG pathway enrichment.

Conclusion: The comparative gene expression analysis shows that ablation of Cyp51 from cholesterol synthesis reduced the hepatic RORCtranscriptional activity stemming from insufficient availability of the natural sterol ligands in vivo.

Search of new glioblastoma stem cell markers by analysis of NCBI GEO datasets, and experimental validation of those markers in glioma cell lines

Marko Vidak1, Damjana Rozman1, Mirjana Liović1, Radovan Komel1

1Institue of Biochemistry, Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia.

Glioblastoma multiforme (GBM) is the most lethal brain tumor. Glioblastoma stem cells are believed to be the reason for the tumor's malignancy and tendency to reappear after a surgical removal. Identification of those stem cells among other cells in the tumor mass would therefore facilitate GBM therapy. Most likely there are different groups of stem cells within a single tumor, each such group expressing its own set of markers. Our goal is to find new markers that would enable targeting of stem cell populations not expressing any of already known markers.

Gene expression profiles of glioblastoma cell lines will be compared with the profiles of non-malignant brain cells from the same dataset. All profiles were obtained by analyzing microarrays and uploaded to NCBI GEO database. The goal is to find genes that are significantly overexpressed in malignant cells in most or all the analyzed datasets. The proteins coded by these genes will be assessed by their function, intracellular location and expression frequency in brain and other tissues. Most interesting candidates, i.e. surface proteins that are not too commonly expressed outside the brain and whose function is related to growth and proliferation, will have their expression investigated in the U87 glioma cell line and the NCH CD133+ purported glioblastoma stem cell line, as well as in a new cell line derived from the U87. Compared to the original U87 line, this new cell line is believed to be enriched in stem-cell like cells, including CD133- stem cells not present in the NCH line.

Mutation analysis of EPO and EPOR genes in two patients with familial erythrocytosis: a novel molecular-genetic diagnostic test for JAK2 negative erythrocytosis

Danijela Vočanec1,2, Tinkara Prijatelj1,2, Tadej Pajič3, Martina Fink3, Irena Preložnik Zupan3, Peter Černelč3, Radovan Komel1, Tanja Kunej2, Nataša Debeljak1

1Institute of Biochemistry, Faculty of Medicine, University of Ljubljana, 2Department of Animal Science, Biotechnical Faculty, University of Ljubljana, 3Clinical department of Haematology, University Medical Centre Ljubljana


Erythrocytosis is heterogeneous group of disorders characterized by the expansion of the erythrocyte compartment including elevated red blood cell (RBC) number, haematocrit, and haemoglobin content in the peripheral blood. Familial erythrocytosis (FE) is a group of rare congenital disorders with various genetic background. Erythropoietin receptor gene (EPOR) mutations are the indicator for primary familial erythrocytosis. Secondary erythrocytosis syndromes are typically associated with a defect in various genes included in oxygen sensing pathway that leads to the increased erythropoietin production [1]. The hormone erythropoietin (EPO) and its receptor (EPOR) are the main regulator of RBC production in the bone marrow. Current diagnostic procedure in Slovenia enables exclusion of JAK2 gene mutations, the cause of polycythaemia vera (PV). Aim of our study was the introduction of new molecular-genetic test for EPO and EPOR genes in JAK2 negative erythrocytosis.


Two related patients with erythrocytosis were negative for JAK2 V617F and JAK2 exon 12 mutations, indicating exclusion of PV. The level of serum erythropoietin was in the reference intervals. Secondary reasons for increased RBC mass (cardiac, pulmonary and endocrine) were excluded. Ensembl database and literature search was performed to detect the most common mutations in EPO and EPOR linked with previously described clinical cases of familial erythrocytosis. Sequence analysis of EPO promoter and 3’ enhancer and EPOR exon 8 was performed.

Primary familial erythrocytosis due to EPOR mutation was excluded by sequence analysis in both patients. So far, 24 mutations in EPOR, located in exon 8, have been associated with erythrocytosis. Exon 8 encodes the C-terminal negative regulatory domain of the protein. Mutations are leading to cytoplasmic truncation of the receptor and loss of the C-terminal negative regulatory domain [2].

However sequence analysis revealed mutation in 3’ enhancer region of the EPO gene in both patients, previously described in blood donors with upper limit haematocrit. The role of erythropoietin in erythrocytosis is indirect and previously had not been linked to the disease.


We have successfully introduced new molecular-genetic test for analysis of the EPO (promoter and 3’ enhancer) and EPOR (exon 8) mutations and implemented it in clinical use. Future recommendation is to complement the diagnostic algorithm with mutational analysis of other genes involved in disease development.


This work was supported by the Slovenian Research Agency (ARRS).


1. Lee G, et al. (2015) Eur J Intern Med 26(5), pp.297-302.

2. Bento C, et al. (2014) Hum Mutat 35(1), pp.15-26.

2006 - University of Ljubljana, Faculty of Medicine, Center for Functional Genomics and Bio-chips.