1. Toward explainable time-series numerical association rule mining : a case study in smart-agricultureIztok Fister, Sancho Salcedo-Sanz, Enrique Alexandre-Cortizo, Damijan Novak, Iztok Fister, Vili Podgorelec, Mario Gorenjak, 2025, izvirni znanstveni članek Opis: This paper defines time-series numerical association rule mining in smart-agriculture applications from an explainable-AI perspective. Two novel explainable methods are presented, along with a newly developed algorithm for time-series numerical association rule mining. Unlike previous approaches, such as fixed interval time-series numerical association, the proposed methods offer enhanced interpretability and an improved data science pipeline by incorporating explainability directly into the software library. The newly developed xNiaARMTS methods are then evaluated through a series of experiments, using real datasets produced from sensors in a smart-agriculture domain. The results obtained using explainable methods within numerical association rule mining in smart-agriculture applications are very positive.
Ključne besede: association rule mining, explainable artificial intelligence, XAI, numerical association rule mining, optimization algorithms
Objavljeno v DKUM: 27.08.2025; Ogledov: 0; Prenosov: 2
 Celotno besedilo (329,69 KB) |
2. Toward precision medicine : molecular biomarkers of response to tofacitinib in inflammatory bowel diseaseAnja Bizjak, Boris Gole, Gregor Jezernik, Uroš Potočnik, Mario Gorenjak, 2025, pregledni znanstveni članek Opis: Ulcerative colitis (UC), a subtype of inflammatory bowel disease (IBD), is a chronic, relapsing inflammatory condition that significantly impairs the patient’s quality of life. While biologics have transformed disease management, a substantial number of patients remain unresponsive or lose efficacy over time. Tofacitinib (TOFA), an oral Janus kinase (JAK) inhibitor, introduces a novel therapeutic class of small-molecule drugs with a unique oral administration route, offering enhanced patient convenience and broader accessibility compared to parenterally administered biologics. As the first oral treatment approved for moderate to severe UC in years, TOFA acts by modulating the JAK/STAT pathway, influencing critical inflammatory mediators such as IL-6, IL-17, and IFN-γ. However, response rates are variable and appear dose-dependent, with up to 60% of patients showing inadequate therapeutic outcomes. This review represents the first comprehensive synthesis focused specifically on biomarkers of TOFA response in UC. Drawing on multi-omics data—epigenomics, transcriptomics, proteomics, and cellular profiling, we highlight emerging predictors of responsiveness, including CpG methylation signatures (e.g., LRPAP1 and FGFR2), transcriptomic regulators (e.g., REG3A and CLDN3), immune and epithelial cell shifts, and the cationic transporter MATE1. TOFA demonstrates a dual mechanism by modulating immune responses while supporting epithelial barrier restoration. Despite being promising, TOFA’s dose-dependent efficacy and interpatient variability underscore the critical need for non-invasive, predictive biomarkers to guide personalized treatment. As the first review of its kind, this work establishes a basis for precision medicine approaches to optimize the clinical utility of TOFA in UC management.
Ključne besede: inflammatory bowel disease, ulcerative colitis, genomics, tofacitinib, transcriptomics, proteomics
Objavljeno v DKUM: 18.08.2025; Ogledov: 0; Prenosov: 12
Celotno besedilo (881,71 KB)
Gradivo ima več datotek! Več... |
3. Rudarjenje asociativnih pravil z evolucijskimi algoritmi : magistrsko deloKatja Božič, 2025, magistrsko delo Opis: V magistrskem delu smo raziskali uporabo evolucijskih algoritmov za rudarjenje asociativnih pravil pri analizi genetskih podatkov na področju bioinformatike. Evolucijske algoritme smo uporabili za odkrivanje znanja iz podatkov, kjer nam asociacijska pravila prikažejo skrite povezave med atributi v podatkovni množici. S pomočjo evolucijskih algoritmov za rudarjenje asociativnih pravil smo na genetskih podatkih zdravih tkiv in tkiv z miomi maternice naredili analizo, s katero smo odkrili neznane povezave med geni. Ugotovili smo, da lahko evolucijske algoritme za rudarjenje asociativnih pravil uporabimo za odkrivanje genetskih vzorcev. Pridobljene podatke bi lahko uporabili za razumevanje molekularnih mehanizmov bolezni, kar bi prispevalo k napredku v diagnostiki in zdravljenju bolezni.
Ključne besede: bioinformatika, evolucijski algoritmi, geni, miomi maternice, rudarjenje asociativnih pravil
Objavljeno v DKUM: 08.05.2025; Ogledov: 0; Prenosov: 36
Celotno besedilo (1,74 MB) |
4. Diferencialno izražanje genov in učinek filtriranja nizko izraženih genov pri analizi RNA sekvenciranjaŠpela Fabčič, 2025, magistrsko delo Opis: Uvod: Sekvenciranje RNA (analiza RNA-seq) se sooča z izzivi obdelave podatkov in razvojem metod za učinkovito shranjevanje, pridobivanje in obdelavo velikih količin podatkov z namenom, da se zmanjša število napak določevanja nukleotidnega zaporedja med drugim tudi na način z odstranitvijo odčitkov nižje kakovosti.
Metode: S pomočjo programskega orodja R smo iz treh različnih podatkovnih baz za tri različne bolezni (Crohnova bolezen, astma, miomi maternice) izvedli RNA-seq analizo v katero smo vključili filtriranje nizko izraženih z metodo CPM, kjer smo določili prag filtriranja nizko izraženih genov na podlagi statističnega izračuna minimuma, maksimuma in povprečja izražanja genov.
Rezultati: V okviru naloge smo ugotovili, da različni pragi filtriranja ne spremenijo števila nabora genov v končnih rezultatih analize RNA-seq. Pri vplivu postavitve različnih pragov filtriranja na nabor prvih 20 najbolj statistično značilno diferencialno izraženih genov je bila izrazita razlika med različno zastavljenimi pragi filtriranja samo v primeru Crohnove bolezni. Pri testiranju, ali se izločeni geni pri danem pragu filtriranja pojavijo kot statistično značilno diferencialno izraženi geni v analizi brez filtriranja, smo ugotovili, da se velik delež izločenih genov, še posebej pri nižjih pragovih filtriranja, še vedno pojavi kot statistično značilen v analizi brez filtriranja.
Razprava in sklep: Na podlagi rezultatov smo ugotovili, da postavitev različnih pragov filtriranja nima posebnega učinka na rezultate diferencialnega izražanja genov.
Ključne besede: genska ekspresija, astma, Crohnova bolezen, miomi maternice
Objavljeno v DKUM: 06.05.2025; Ogledov: 0; Prenosov: 14
Celotno besedilo (1,39 MB) |
5. Genetic background of cattle temperament : a short reviewJože Smolinger, Mario Gorenjak, Dejan Škorjanc, 2024, pregledni znanstveni članek Opis: Animal temperament describes behavioural differences between individuals that are consistent over time and across different circumstances. Knowledge of the animal's temperament has a major effect on the safety of handling and caring for the animals as well as on the adaptation of the animals to changing rearing conditions. To understand animal temperament, we need to know not only the genetic basis of temperament, but also the influence of the environment on its expression. Similarly the temperament of dairy cows can be defined as the animal's response to environmental or social stimuli. In this review article, chromosomes with genomic regions containing QTLs, genes and candidate genes responsible for the expression of temperament traits in cattle are presented. Knowledge of the genetic background of temperament expression in cattle and its variability in these traits allows temperament to be included in the selection index.
Ključne besede: cattle temperament genetics, QTL, SNP, heritability, serotonin, dopamine
Objavljeno v DKUM: 23.04.2025; Ogledov: 0; Prenosov: 1
 Povezava na datoteko |
6. MFUM-BrTNBC-1, a newly established patient-derived triple-negative breast cancer cell line : molecular characterisation, genetic stability, and comprehensive comparison with commercial breast cancer cell linesKristijan Skok, Lidija Gradišnik, Helena Sabina Čelešnik, Marko Milojević, Uroš Potočnik, Gregor Jezernik, Mario Gorenjak, Monika Sobočan, Iztok Takač, Rajko Kavalar, Uroš Maver, 2022, izvirni znanstveni članek Opis: Triple-negative breast cancer (TNBC) is a breast cancer (BC) subtype that accounts for
approximately 15–20% of all BC cases. Cancer cell lines (CLs) provide an efficient way to model the
disease. We have recently isolated a patient-derived triple-negative BC CL MFUM-BrTNBC-1 and
performed a detailed morphological and molecular characterisation and a comprehensive comparison
with three commercial BC CLs (MCF-7, MDA-MB-231, MDA-MB-453). Light and fluorescence
microscopy were used for morphological studies; immunocytochemical staining for hormone receptor,
p53 and Ki67 status; RNA sequencing, qRT-PCR and STR analysis for molecular characterisation; and
biomedical image analysis for comparative phenotypical analysis. The patient tissue-derived MFUMBrTNBC-1 maintained the primary triple-negative receptor status. STR analysis showed a stable and
unique STR profile up to the 6th passage. MFUM-BrTNBC-1 expressed EMT transition markers and
displayed changes in several cancer-related pathways (MAPK, Wnt and PI3K signalling; nucleotide
excision repair; and SWI/SNF chromatin remodelling). Morphologically, MFUM-BrTNBC-1 differed
from the commercial TNBC CL MDA-MB-231. The advantages of MFUM-BrTNBC-1 are its isolation
from a primary tumour, rather than a metastatic site; good growth characteristics; phenotype identical
to primary tissue; complete records of origin; a unique identifier; complete, unique STR profile;
quantifiable morphological properties; and genetic stability up to (at least) the 6th passage.
Ključne besede: hormonal receptors, MFUM-BrTNBC-1, MCF-7, MDA-MB-231, MDA-MB-453
Objavljeno v DKUM: 10.04.2025; Ogledov: 0; Prenosov: 8
Celotno besedilo (5,33 MB)
Gradivo ima več datotek! Več... |
7. Impaired neurodevelopmental genes in Slovenian autistic children elucidate the comorbidity of autism with other developmental disordersDanijela Krgović, Mario Gorenjak, Nika Rihar, Iva Opalič, Špela Stangler Herodež, Hojka Gregorič Kumperščak, Peter Dovč, Nadja Kokalj-Vokač, 2022, izvirni znanstveni članek Opis: Autism spectrum disorders (ASD) represent a phenotypically heterogeneous group of patients that strongly intertwine with other neurodevelopmental disorders (NDDs), with genetics playing a significant role in their etiology. Whole exome sequencing (WES) has become predominant in molecular diagnostics for ASD by considerably increasing the diagnostic yield. However, the proportion of undiagnosed patients still remains high due to complex clinical presentation, reduced penetrance, and lack of segregation analysis or clinical information. Thus, reverse phenotyping, where we first identified a possible genetic cause and then determine its clinical relevance, has been shown to be a more efficient approach. WES was performed on 147 Slovenian pediatric patients with suspected ASD. Data analysis was focused on identifying ultrarare or “single event” variants in ASD-associated genes and further expanded to NDD-associated genes. Protein function and gene prioritization were performed on detected clinically relevant variants to determine their role in ASD etiology and phenotype. Reverse phenotyping revealed a pathogenic or likely pathogenic variant in ASD-associated genes in 20.4% of patients, with subsequent segregation analysis indicating that 14 were de novo variants and 1 was presumed compound heterozygous. The diagnostic yield was further increased by 2.7% by the analysis of ultrarare or “single event” variants in all NDD-associated genes. Protein function analysis established that genes in which variants of unknown significance (VUS) were detected were predominantly the cause of intellectual disability (ID), and in most cases, features of ASD as well. Using such an approach, variants in rarely described ASD-associated genes, such as SIN3B, NR4A2, and GRIA1, were detected. By expanding the analysis to include functionally similar NDD genes, variants in KCNK9, GNE, and other genes were identified. These would probably have been missed by classic genotype–phenotype analysis. Our study thus demonstrates that in patients with ASD, analysis of ultrarare or “single event” variants obtained using WES with the inclusion of functionally similar genes and reverse phenotyping obtained a higher diagnostic yield despite limited clinical data. The present study also demonstrates that most of the causative genes in our cohort were involved in the syndromic form of ASD and confirms their comorbidity with other developmental disorders.
Ključne besede: reverse phenotyping, single event variants, NDD-associated genes, GRIA1 gene, NR4A2 gene, SIN3B gene, autism, child
Objavljeno v DKUM: 12.12.2024; Ogledov: 0; Prenosov: 8
Celotno besedilo (6,36 MB)
Gradivo ima več datotek! Več... |
8. Single-cell transcriptomic and targeted genomic profiling adjusted for inflammation and therapy bias reveal CRTAM and PLCB1 as novel hub genes for anti-tumor necrosis factor alpha therapy response in Crohn’s diseaseMario Gorenjak, Boris Gole, Larisa Goričan, Gregor Jezernik, Uršula Prosenc Zmrzljak, Cvetka Pernat Drobež, Pavel Skok, Uroš Potočnik, 2024, izvirni znanstveni članek Opis: The lack of reliable biomarkers in response to anti-TNFα biologicals hinders
personalized therapy for Crohn’s disease (CD) patients. The motivation behind our study is to shift
the paradigm of anti-TNFα biomarker discovery toward specific immune cell sub-populations using
single-cell RNA sequencing and an innovative approach designed to uncover PBMCs gene expression
signals, which may be masked due to the treatment or ongoing inflammation; Methods: The singlecell
RNA sequencing was performed on PBMC samples from CD patients either naïve to biological
therapy, in remission while on adalimumab, or while on ustekinumab but previously non-responsive
to adalimumab. Sieves for stringent downstream gene selection consisted of gene ontology and
independent cohort genomic profiling. Replication and meta-analyses were performed using publicly
available raw RNA sequencing files of sorted immune cells and an association analysis summary.
Machine learning, Mendelian randomization, and oligogenic risk score methods were deployed to
validate DEGs highly relevant to anti-TNFα therapy response; Results: This study found PLCB1 in
CD4+ T cells and CRTAM in double-negative T cells, which met the stringent statistical thresholds
throughout the analyses. An additional assessment proved causal inference of both genes in response
to anti-TNFα therapy; Conclusions: This study, jointly with an innovative design, uncovered
novel candidate genes in the anti-TNFα response landscape of CD, potentially obscured by therapy
or inflammation.
Ključne besede: inflammatory bowel diseases, Crohn’s disease, tumor necrosis factor alpha, adalimumab, single-cell gene expression analysis
Objavljeno v DKUM: 10.12.2024; Ogledov: 0; Prenosov: 13
Celotno besedilo (7,36 MB)
Gradivo ima več datotek! Več... |
9. Discovery of novel biomarkers with extended non-coding RNA interactor networks from genetic and protein biomarkersGregor Jezernik, Damjan Glavač, Pavel Skok, Martina Krušič, Uroš Potočnik, Mario Gorenjak, 2024, izvirni znanstveni članek Opis: Curated online interaction databases and gene ontology tools have streamlined the analysis of highly complex gene/protein networks. However, understanding of disease pathogenesis has gradually shifted from a protein-based core to complex interactive networks where non-coding RNA (ncRNA) is thought to play an essential role. As current gene ontology is based predominantly on protein-level information, there is a growing need to analyze networks with ncRNA. In this study, we propose a gene ontology workflow integrating ncRNA using the NPInter V5.0 database. To validate the proposed workflow, we analyzed our previously published curated biomarker datasets for hidden disease susceptibility processes and pharmacogenomics. Our results show a novel involvement of melanogenesis in psoriasis response to biological drugs in general. Hyperpigmentation has been previously observed in psoriasis following treatment with currently indicated biological drugs, thus calling attention to melanogenesis research as a response biomarker in psoriasis. Moreover, our proposed workflow highlights the need to critically evaluate computed ncRNA interactions within databases and a demand for gene ontology analysis of large miRNA blocks.
Ključne besede: gene ontology, non-coding RNA, ncRNA, disease pathogenesis, gene networks, protein networks, tools, psoriasis
Objavljeno v DKUM: 06.12.2024; Ogledov: 0; Prenosov: 13
Celotno besedilo (9,14 MB)
Gradivo ima več datotek! Več... |
10. Meta-analytic comparison of global RNA transcriptomes of acute and chronic myeloid leukemia cells reveals novel gene candidates governing myeloid malignanciesStaša Jurgec, Gregor Jezernik, Mario Gorenjak, Tomaž Büdefeld, Uroš Potočnik, 2022, izvirni znanstveni članek Opis: Despite advances in the understanding of genetic risk factors and molecular mechanisms underlying acute myeloid leukemia (AML) and chronic myeloid leukemia (CML), clinical outcomes of current therapies in terms of disease relapse and mortality rate pose a great economic and social burden. To overcome this, the identification of new molecular prognostic biomarkers and pharmacological targets is crucial. Recent studies have suggested that AML and CML may share common pathogenic mechanisms and cellular substrates. To this end, in the present study, global transcriptome profiles of AML and CML at the molecular and cellular level were directly compared using a combination of meta-analysis and modern statistics, and novel candidate genes and specific biological processes associated with the pathogenesis of AML and CML were characterized. Our study significantly improves our current understanding of myeloid leukemia and will help develop new therapeutic targets and biomarkers for disease progression, management and treatment response.
Ključne besede: AML, CML, meta-analysis, lincRNA, spliceosome
Objavljeno v DKUM: 05.12.2024; Ogledov: 0; Prenosov: 10
Celotno besedilo (1,04 MB)
Gradivo ima več datotek! Več... |
