Peter Juvan

University of Ljubljana, Lubliano, Ljubljana, Slovenia

Are you Peter Juvan?

Claim your profile

Publications (25)95.5 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: Abstract 1. Nucleotide analogues comprise an important class of drugs used in treatment of viral infections but also cancer. These drugs affect the structural integrity of DNA and activate different pathways and processes in the cell and may directly or indirectly influence the drug metabolizing system. Adefovir dipivoxil (AD) and tenofovir disoproxil (TD) are nucleotide analogues approved for the treatment of chronic hepatitis B and/or HIV/AIDS infection. 2. To evaluate the risk of their drug-drug interactions on the level of drug metabolism, an effect of both compounds on cytochromes P450 expression was studied using cDNA microarrays, real-time RT-PCR and immunoblotting. Mice were given intraperitoneally 25 mg/kg of AD or TD, respectively. As a positive control, a combination of prototypic cytochromes P450 (CYP) inducers, phenobarbital and β-naphthoflavone was chosen. 3. The data obtained showed a significant CYP induction in the positive control group, but no clinically significant induction of CYP genes by AD or TD was observed. Our results support the evidence of safety of AD and TD with respect to drug-drug interactions based on enzyme induction. These findings are important as a plethora of new antivirals of different types are being tested and introduced to clinical practice, mostly to be used in combinations.
    Xenobiotica 03/2014; · 1.98 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We examined the genotype-phenotype interactions of Cyp51+/- mice carrying one functional allele of lanosterol 14α-demethylase from cholesterol biosynthesis. No distinct developmental or morphological abnormalities were observed by routine visual inspection of Cyp51+/- and Cyp51+/+ mice and fertility was similar. We further collected a large data-set from female and male Cyp51+/- mice and controls fed for 16 weeks with three diets and applied linear regression modeling. We used 3 predictor variables (genotype, sex, diet), and 39 response variables corresponding to the organ characteristics (7), plasma parameters (7), and hepatic gene expression (25). We observed significant differences between Cyp51+/- and wild-type mice in organ characteristics and blood lipid profile. Hepatomegaly was observed in Cyp51+/- males, together with elevated total and low-density lipoprotein cholesterol. Cyp51+/- females fed high-fat, high-cholesterol diet were leaner and had elevated plasma corticosterone compared to controls. We observed elevated hepatocyte apoptosis, mitosis and lipid infiltration in heterozygous knockouts of both sexes. The Cyp51+/- females had a modified lipid storage homeostasis protecting them from weight-gain when fed high-fat high-cholesterol diet. Malfunction of one Cyp51 allele therefore initiates disease pathways towards cholesterol-linked liver pathologies and sex-dependent response to dietary challenge.
    PLoS ONE 01/2014; 9(11):e112787. · 3.53 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The aim of our study was to determine whether there are any differences in the cumulus cell gene expression profile of mature oocytes derived from modified natural IVF and controlled ovarian hyperstimulation cycles and if these changes could help us understand why modified natural IVF has lower success rates. Cumulus cells surrounding mature oocytes that developed to morulae or blastocysts on day 5 after oocyte retrieval were submitted to microarray analysis. The obtained data were then validated using quantitative real-time PCR. There were 66 differentially expressed genes between cumulus cells of modified natural IVF and controlled ovarian hyperstimulation cycles. Gene ontology analysis revealed the oxidation-reduction process, glutathione metabolic process, xenobiotic metabolic process and gene expression were significantly enriched biological processes in MNIVF cycles. Among differentially expressed genes we observed a large group of small nucleolar RNA's whose role in folliculogenesis has not yet been established. The increased expression of genes involved in the oxidation-reduction process probably points to hypoxic conditions in modified natural IVF cycles. This finding opens up new perspectives for the establishment of the potential role that oxidation-reduction processes have in determining success rates of modified natural IVF.
    Journal of Assisted Reproduction and Genetics 11/2013; · 1.82 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: A universal biomarker panel with the potential to predict high-risk pregnancies or adverse pregnancy outcome does not exist. Transcriptome analysis is a powerful tool to capture differentially expressed genes (DEG), which can be used as biomarker-diagnostic-predictive tool for various conditions in prenatal setting. In search of biomarker set for predicting high-risk pregnancies, we performed global expression profiling to find DEG in Ts21. Subsequently, we performed targeted validation and diagnostic performance evaluation on a larger group of case and control samples. Initially, transcriptomic profiles of 10 cultivated amniocyte samples with Ts21 and 9 with normal euploid constitution were determined using expression microarrays. Datasets from Ts21 transcriptomic studies from GEO repository were incorporated. DEG were discovered using linear regression modelling and validated using RT-PCR quantification on an independent sample of 16 cases with Ts21 and 32 controls. The classification performance of Ts21 status based on expression profiling was performed using supervised machine learning algorithm and evaluated using a leave-one-out cross validation approach. Global gene expression profiling has revealed significant expression changes between normal and Ts21 samples, which in combination with data from previously performed Ts21 transcriptomic studies, were used to generate a multi-gene biomarker for Ts21, comprising of 9 gene expression profiles. In addition to biomarker's high performance in discriminating samples from global expression profiling, we were also able to show its discriminatory performance on a larger sample set 2, validated using RT-PCR experiment (AUC=0.97), while its performance on data from previously published studies reached discriminatory AUC values of 1.00. Our results show that transcriptomic changes might potentially be used to discriminate trisomy of chromosome 21 in the prenatal setting. As expressional alterations reflect both, causal and reactive cellular mechanisms, transcriptomic changes may thus have future potential in the diagnosis of a wide array of heterogeneous diseases that result from genetic disturbances.
    PLoS ONE 01/2013; 8(9):e74184. · 3.53 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Human recombinant erythropoietin (rHuEpo) that is used for the treatment of the chemotherapy-induced anaemia in cancer patients was shown to cause detrimental effects on the course of disease due to increased adverse events inflicting patient's survival, potentially related to rHuEpo-induced cancer progression. In this study, we elucidate the effect of rHuEpo administration on breast cancer cell proliferation and gene expression after cisplatin (cDDP) induced cytotoxicity. Two breast carcinoma models, MCF-7 and MDA-MB-231 cell lines, were used differing in oestrogen (ER) and progesterone (PR) receptors and p53 status. Cells were cultured with or without rHuEpo for 24 h or 9 weeks and their growth characteristics after cDDP treatment were assessed together with expression of genes involved in the p53-signaling pathway. Short-term exposure of breast cancer cells to rHuEpo lowers their proliferation and reduces cDDP cytotoxic potency. In contrast, long-term exposure of MCF-7 cells to rHuEpo increases proliferation and predisposes MCF-7 cells to cDDP cytotoxicity, but has no effect on MDA-MB-231 cells. MDA-MB-231 cells show altered level of ERK phosphorylation, indicating involvement of MAPK signalling pathway. Gene expression analysis of p53-dependent genes and bcl-2 gene family members confirmed differences between long and short-term rHuEpo effects, indicating the most prominent changes in BCL2 and BAD expression. Proliferation and survival characteristics of MCF-7 cells are reversely modulated by the length of the rHuEpo exposure. On the other hand, MDA-MB-231 cells are almost irresponsive to long-term rHuEpo, supposedly due to the mutated p53 and ER(+)/PR(-) status. The p53 and ER/PR status may predict tumour response on rHuEpo and cDDP treatment.
    Radiology and Oncology 09/2012; 46(3):213-25. · 1.60 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The essential role of the Crem gene in normal sperm development is widely accepted and is confirmed by azoospermia in male mice lacking the Crem gene. The exact number of genes affected by Crem absence is not known, however a large difference has been observed recently between the estimated number of differentially expressed genes found in Crem knock-out (KO) mice compared to the number of gene loci bound by CREM. We therefore re-examined global gene expression in male mice lacking the Crem gene using whole genome transcriptome analysis with Affymetrix microarrays and compared the lists of differentially expressed genes from Crem-/- mice to a dataset of genes where binding of CREM was determined by Chip-seq. We determined the global effect of CREM on spermatogenesis as well as distinguished between primary and secondary effects of the CREM absence. We demonstrated that the absence of Crem deregulates over 4700 genes in KO testis. Among them are 101 genes associated with spermatogenesis 41 of which are bound by CREM and are deregulated in Crem KO testis. Absence of several of these genes in mouse models has proven their importance for normal spermatogenesis and male fertility. Our study showed that the absence of Crem plays a more important role on different aspects of spermatogenesis as estimated previously, with its impact ranging from apoptosis induction to deregulation of major circadian clock genes, steroidogenesis and the cell-cell junction dynamics. Several new genes important for normal spermatogenesis and fertility are down-regulated in KO testis and are therefore possible novel targets of CREM.
    PLoS ONE 01/2012; 7(2):e31798. · 3.53 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: In in vitro fertilization (IVF) cycles controlled ovarian hyperstimulation (COH) is established by gonadotropins in combination with gonadotropin-releasing hormone (GnRH) agonists or antagonists, to prevent premature luteinizing hormone (LH) surge. The aim of our study was to improve the understanding of gene expression profile of cumulus cells (CC) in terms of ovarian stimulation protocol and oocyte maturity. We applied Affymetrix gene expression profiling in CC of oocytes at different maturation stages using either GnRH agonists or GnRH antagonists. Two analyses were performed: the first involved CC of immature metaphase I (MI) and mature metaphase II (MII) oocytes where 359 genes were differentially expressed, and the second involved the two GnRH analogues where no differentially expressed genes were observed at the entire transcriptome level. A further analysis of 359 differentially genes was performed, focusing on anti-Müllerian hormone receptor 2 (AMHR2), follicle stimulating hormone receptor (FSHR), vascular endothelial growth factor C (VEGFC) and serine protease inhibitor E2 (SERPINE2). Among other differentially expressed genes we observed a marked number of new genes connected to cell adhesion and neurotransmitters such as dopamine, glycine and γ-Aminobutyric acid (GABA). No differential expression in CC between the two GnRH analogues supports the findings of clinical studies where no significant difference in live birth rates between both GnRH analogues has been proven.
    PLoS ONE 01/2012; 7(10):e47106. · 3.53 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: With particular emphasis on interactions between cholesterol homeostasis and drug metabolism we investigate the transcriptome of human primary hepatocytes treated by two commonly prescribed cholesterol lowering drugs atorvastatin and rosuvastatin and by rifampicin that serves as an outgroup as well as a model substance for induction of nuclear pregnane X receptor. Hepatocytes from human donors have been treated with rosuvastatin, atorvastatin, and rifampicin for 12, 24, and 48 h. Expression profiling with cholesterol and drug metabolism enriched low density Steroltalk cDNA and whole genome Affymetrix HG-U133 Plus 2.0 arrays has been applied. Differential expression (DE) of genes and gene set enrichment analysis of KEGG pathways were performed. Lists of differentially expressed genes and gene sets were cross-compared. Selected genes were confirmed by quantitative real-time PCR. Statins lead to: (a) upregulation of cholesterol-related genes indicating an increased LDL uptake and storage of esterified cholesterol, elevated bile acid/drug export and lower capacity to form HDL; (b) perturbation of genes in glucose and fatty acid homeostasis, influencing acetyl-CoA pools, promoting gluconeogenesis and glucose export; (c) elevated expression of ADIPOR2 suggesting increased sensitivity to adiponectin; (d) perturbations in genes of lipoprotein particle formation, differently for each statin; (e) perturbed expression of many metabolic genes that are directly controlled by nuclear receptors constitutive androstan and/or pregnane X. These data provide a novel global insight into hepatic effects of statins, offering biochemical explanations for higher blood glucose in statin-treated patients, and for drug-induced secondary fatty liver disease.
    Pharmacogenetics and Genomics 08/2011; 21(11):741-50. · 3.61 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Aldehydes are ubiquitous natural constituents of foods, water and beverages. Dietary intake represents the greatest source of exposure to acrolein and related aldehydes. Oral acrolein induces dyslipidemia acutely and chronically increases atherosclerosis in mice, yet the mechanisms are unknown. Because lipid synthesis and trafficking are largely under hepatic control, we examined hepatic genes in murine models of acute and chronic oral acrolein exposure. Changes in hepatic gene expression were examined using a Steroltalk microarray. Acute acrolein feeding modified plasma and hepatic proteins and increased plasma triglycerides within 15  min. By 6  h, acrolein altered hepatic gene expression including Insig1, Insig2 and Hmgcr genes and stimulated an acute-phase response (APR) with up-regulation of serum amyloid A genes (Saa) and systemic hypoalbuminemia. To test if decreased HMG-CoA reductase activity could modify acrolein-induced dyslipidemia or the APR, mice were pretreated with simvastatin. Statin treatment, however, did not alter acrolein-induced dyslipidemia or hypoalbuminemia associated with an APR. Few hepatic genes were dysregulated by chronic acrolein feeding in apoE-null mice. These studies confirmed that acute acrolein exposure altered expression of hepatic genes involved with lipid synthesis and trafficking and APR, and thus, indicated a hepatic locus of acrolein-induced dyslipidemia and APR that was independent of HMG CoA-reductase. Dietary intake of acrolein could contribute to cardiovascular disease risk by disturbing hepatic function.
    Molecular Nutrition & Food Research 08/2011; 55(9):1411-22. · 4.31 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Divergently selected Lean and Fat mouse lines represent unique models for a polygenic form of resistance and susceptibility to obesity development. Previous research on these lines focused mainly on obesity-susceptible factors in the Fat line. This study aimed to examine the molecular basis of obesity-resistant mechanisms in the Lean line by analyzing various fat depots and organs, the liver transcriptome of selected metabolic pathways, plasma and lipid homeostasis and expression of selected skeletal muscle genes. Expression profiling using our custom Steroltalk v2 microarray demonstrated that Lean mice exhibit a higher hepatic expression of cholesterol biosynthesis genes compared to the Fat line, although this was not reflected in elevation of total plasma or liver cholesterol. However, FPLC analysis showed that protective HDL cholesterol was elevated in Lean mice. A significant difference between the strains was also found in bile acid metabolism. Lean mice had a higher expression of Cyp8b1, a regulatory enzyme of bile acid synthesis, and the Abcb11 bile acid transporter gene responsible for export of acids to the bile. Additionally, a higher content of blood circulating bile acids was observed in Lean mice. Elevated HDL and upregulation of some bile acids synthesis and transport genes suggests enhanced reverse cholesterol transport in the Lean line--the flux of cholesterol out of the body is higher which is compensated by upregulation of endogenous cholesterol biosynthesis. Increased skeletal muscle Il6 and Dio2 mRNA levels as well as increased activity of muscle succinic acid dehydrogenase (SDH) in the Lean mice demonstrates for the first time that changes in muscle energy metabolism play important role in the Lean line phenotype determination and corroborate our previous findings of increased physical activity and thermogenesis in this line. Finally, differential expression of Abcb11 and Dio2 identifies novel strong positional candidate genes as they map within the quantitative trait loci (QTL) regions detected previously in crosses between the Lean and Fat mice. We identified novel candidate molecular targets and metabolic changes which can at least in part explain resistance to obesity development in the Lean line. The major difference between the Lean and Fat mice was in increased liver cholesterol biosynthesis gene mRNA expression, bile acid metabolism and changes in selected muscle genes' expression in the Lean line. The liver Abcb11 and muscle Dio2 were identified as novel positional candidate genes to explain part of the phenotypic difference between the Lean and Fat lines.
    BMC Genomics 02/2011; 12:96. · 4.40 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Aldehydes such as acrolein are ubiquitous pollutants present in automobile exhaust, cigarette, wood, and coal smoke. Such aldehydes are also constituents of several food substances and are present in drinking water, irrigation canals, and effluents from manufacturing plants. Oral intake represents the most significant source of exposure to acrolein and related aldehydes. To study the effects of short-term oral exposure to acrolein on lipoprotein levels and metabolism, adult mice were gavage-fed 0.1 to 5 mg acrolein/kg bwt and changes in plasma lipoproteins were assessed. Changes in hepatic gene expression related to lipid metabolism and cytokines were examined by qRT-PCR analysis. Acrolein feeding did not affect body weight, blood urea nitrogen, plasma creatinine, electrolytes, cytokines or liver enzymes, but increased plasma cholesterol and triglycerides. Similar results were obtained with apoE-null mice. Plasma lipoproteins from acrolein-fed mice showed altered electrophoretic mobility on agarose gels. Chromatographic analysis revealed elevated VLDL cholesterol, phospholipids, and triglycerides levels with little change in LDL or HDL. NMR analysis indicated shifts from small to large VLDL and from large to medium-small LDL with no change in the size of HDL particles. Increased plasma VLDL was associated with a significant decrease in post-heparin plasma hepatic lipase activity and a decrease in hepatic expression of hepatic lipase. These observations suggest that oral exposure to acrolein could induce or exacerbate systemic dyslipidemia and thereby contribute to cardiovascular disease risk.
    Toxicology and Applied Pharmacology 02/2010; 243(1):1-12. · 3.98 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The nuclear receptors CAR (constitutive androstane receptor) and PXR (pregnane X receptor) mediate the effects of phenobarbital on gene transcription. To investigate the relative contribution of these nuclear receptors to the expression of specific genes we studied the effect of phenobarbital in livers of wild type, CAR(-/-), PXR(-/-) and CAR/PXR(-/-) knockout mice. Spotted Steroltalk v1 cDNA arrays were applied containing probes for genes involved in drug metabolism, sterol biosynthesis, steroid synthesis/transport and heme synthesis. In the absence of CAR and PXR, phenobarbital unexpectedly induced mRNAs of several nuclear receptors, including PPARalpha and its target genes Cyp4a10 and Cyp4a14. Interestingly, in primary cultures of hepatocytes isolated from CAR/PXR(-/-) knockout mice, phenobarbital increased HNF-4alpha levels. In further experiments in these hepatocyte cultures we provide evidence that phenobarbital directly induces transcription of the PPARalpha gene via its HNF-4alpha response element, and indirectly by lack of inhibitory crosstalk of AMPK, CAR and PXR with HNF-4alpha. Our results provide further insight into CAR and PXR-independent effects of phenobarbital and the crosstalk between different nuclear receptor signaling pathways.
    Molecular Pharmaceutics 09/2009; 6(5):1573-81. · 4.57 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Cholesterol homeostasis and xenobiotic metabolism are complex biological processes, which are difficult to study with traditional methods. Deciphering complex regulation and response of these two processes to different factors is crucial also for understanding of disease development. Systems biology tools as are microarrays can importantly contribute to this knowledge and can also discover novel interactions between the two processes. We have developed a low density Sterolgene v0 cDNA microarray dedicated to studies of cholesterol homeostasis and drug metabolism in the mouse. To illustrate its performance, we have analyzed mouse liver samples from studies focused on regulation of cholesterol homeostasis and drug metabolism by diet, drugs and inflammation. We observed down-regulation of cholesterol biosynthesis during fasting and high-cholesterol diet and subsequent up-regulation by inflammation. Drug metabolism was down-regulated by fasting and inflammation, but up-regulated by phenobarbital treatment and high-cholesterol diet. Additionally, the performance of the Sterolgene v0 was compared to the two commercial high density microarray platforms: the Agilent cDNA (G4104A) and the Affymetrix MOE430A GeneChip. We hybridized identical RNA samples to the commercial microarrays and showed that the performance of Sterolgene is comparable to commercial arrays in terms of detection of changes in cholesterol homeostasis and drug metabolism. Using the Sterolgene v0 microarray we were able to detect important changes in cholesterol homeostasis and drug metabolism caused by diet, drugs and inflammation. Together with its next generations the Sterolgene microarrays represent original and dedicated tools enabling focused and cost effective studies of cholesterol homeostasis and drug metabolism. These microarrays have the potential of being further developed into screening or diagnostic tools.
    BMC Genomics 02/2008; 9:76. · 4.40 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: It has long been demonstrated that the level of cholesterol in cells regulates the cholesterol biosynthesis through SREBF transcription factors, but lately it has been shown that other factors are also important. To study the system we employed Bayesian network inference and combined it with mathematical modeling and simulation. We constructed a mathemat-ical model of cholesterol biosynthesis and studied its properties through simulation. We measured transcriptional changes of cholesterogenic genes using the Steroltalk microarray and treated human hepatocyte samples. We employed Bayesian inference to identify gene-to-gene interactions from both microarray measurements and simulated data. The inferred networks show that the expression of cholesterogenic genes can be predicted from the expression of 4 key genes, one of them being SREBF2. Networks also indicate a strong interaction between SREBF2 and CYP51A1, but not between SREBF2 and HMGCR, the rate-limiting enzyme of cholesterol biosynthesis. The expression of HMGCR seems to be regulated by other factor(s). Computer simulations of the mathematical model of cholesterol biosynthesis exposed that a large number of perturbations of the system is critical for identification of gene-to-gene interactions, and that dif-ferences between human individuals (biological variability) and measurement noise (technical variability) pose a seri-ous problem for their automatic inference from DNA microarray data.
    Acta Chim. Slov. 01/2008; 55:396-407.
  • [Show abstract] [Hide abstract]
    ABSTRACT: GenePath is an automated system for reasoning about genetic networks, wherein a set of genes have various influences on one another and on a biological outcome. It acts on a set of experiments in which genes are knocked-out or overexpressed, and the outcome of interest is evaluated. Implemented in Prolog, GenePath uses abductive inference to elucidate network constraints based on background knowledge and experimental results. Two uses of the system are demonstrated: synthesis of a consistent network from abduced constraints, and qualitative reasoning-based approach that generates a set of networks consistent with the data. In practice, as illustrated by an example on aggregation of a soil amoeba Dictyostelium discoideum, a combination of constraint satisfaction and qualitative reasoning produces a small set of plausible networks.
    Computational Discovery of Scientific Knowledge, Introduction, Techniques, and Applications in Environmental and Life Sciences; 01/2007
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: GenePath is a web-based application for the analysis of mutant-based experiments and synthesis of genetic networks. Here, we introduce GenePath and describe a number of new approaches, including conflict resolution, handling cyclic pathways, confidence level assignment, what-if analysis and new experiment proposal. We illustrate the key concepts using data from a study of adhesion genes in Dictyostelium discoideum and show that GenePath discovered genetic interactions that were ignored in the original publication. GenePath is available at http://www.genepath.org/genepath2.
    Nucleic Acids Research 08/2005; 33(Web Server issue):W749-52. · 8.81 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Classical epistasis analysis can determine the order of function of genes in pathways using morphological, biochemical and other phenotypes. It requires knowledge of the pathway's phenotypic output and a variety of experimental expertise and so is unsuitable for genome-scale analysis. Here we used microarray profiles of mutants as phenotypes for epistasis analysis. Considering genes that regulate activity of protein kinase A in Dictyostelium, we identified known and unknown epistatic relationships and reconstructed a genetic network with microarray phenotypes alone. This work shows that microarray data can provide a uniform, quantitative tool for large-scale genetic network analysis.
    Nature Genetics 06/2005; 37(5):471-7. · 35.21 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Genetic networks are often used in the analysis of biological phenomena. In classical genetics, they are constructed manually from experimental data on mutants. The field lacks formalism to guide such analysis, and accounting for all the data becomes complicated when large amounts of data are considered. We have developed GenePath, an intelligent assistant that automates the analysis of genetic data. GenePath employs expert-defined patterns to uncover gene relations from the data, and uses these relations as constraints in the search for a plausible genetic network. GenePath formalizes genetic data analysis, facilitates the consideration of all the available data in a consistent manner, and the examination of the large number of possible consequences of planned experiments. It also provides an explanation mechanism that traces every finding to the pertinent data. GenePath can be accessed at http://genepath.org. Supplementary material is available at http://genepath.org/bi-.supp.
    Bioinformatics 03/2003; 19(3):383-9. · 5.32 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: A genetic network is a formalism that is often used in biology to represent causalities and reason about biological phenomena related to genetic regulation. We present GenePath, a computer-based system that supports the inference of genetic networks from a set of genetic experiments. Implemented in Prolog, GenePath uses abductive inference to elucidate network constraints based on background knowledge and experimental results. Additionally, it can propose genetic experiments that may further refine the discovered network and establish relations between genes that could not be related based on the original experimental data. We illustrate GenePath's approach and utility on analysis of data on aggregation and sporulation of the soil amoeba Dictyostelium discoideum.
    Artificial Intelligence in Medicine 01/2003; 29(1-2):107-30. · 1.36 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Genetic pathways and networks are often used in the analysis of biological phenomena. They are constructed by determining the effect of mutations in different genes on the investigated phenomenon and by determining the relationships between the mutated genes. We describe a software tool named GenePath (http://genepath.org/) which serves as an assistant for genetic pathway discovery. The pathways and genetic networks are constructed in two steps. In the first step, genetic data are matched to a set of inference patterns of the type "If a combination of experiments is found in the data, THEN a relation between genes and phenotypic states can be assumed". The discovered relations impose constraints over the plausible genetic network, which is constructed in the second step such that all the constraints are satisfied. In addition to automation of the network discovery process, GenePath provides an 'explanation', the tracing of any network relation back to the pertinent experimental data. GenePath was trained on a genetic pathway that regulates the transition from growth to development in the social amoeba Dictyostelium discoideum and tested successfully on several other problems, demonstrating its utility as a general-purpose intelligent assistant to biologists in their explorative analysis of genetic data.
    01/2002; 3.

Publication Stats

237 Citations
95.50 Total Impact Points

Institutions

  • 2000–2013
    • University of Ljubljana
      • • Institute of Biochemistry
      • • Faculty of Medicine
      • • Center for Functional Genomics and Bio-Chips (CFGBC)
      • • Faculty of Computer and Information Science
      Lubliano, Ljubljana, Slovenia
  • 2005
    • Baylor College of Medicine
      • Department of Molecular & Human Genetics
      Houston, TX, United States