A5072943068- Data Mining Algorithms and Applications
- Genomics and Phylogenetic Studies
- Bioinformatics and Genomic Networks
- Machine Learning in Bioinformatics
- Gene expression and cancer classification
- Biomedical Text Mining and Ontologies
- Computational Drug Discovery Methods
- Rough Sets and Fuzzy Logic
- Nutrition, Genetics, and Disease
- Complex Network Analysis Techniques
- Machine Learning and Data Classification
- Data Management and Algorithms
- Complex Systems and Time Series Analysis
- State Capitalism and Financial Governance
- Recommender Systems and Techniques
- Global Financial Crisis and Policies
- Opinion Dynamics and Social Influence
- Microbial Natural Products and Biosynthesis
- Radiation Shielding Materials Analysis
- Heart Rate Variability and Autonomic Control
- Graphite, nuclear technology, radiation studies
- Identification and Quantification in Food
- Nuclear and radioactivity studies
- Text and Document Classification Technologies
- Aquaculture Nutrition and Growth
Rudjer Boskovic Institute
2015-2025
Laureate Education
2008-2017
University of Zagreb
2010
Croatian Veterinary Institute
2008
Outcomes of high-throughput biological experiments are typically interpreted by statistical testing for enriched gene functional categories defined the Gene Ontology (GO). The resulting lists GO terms may be large and highly redundant, thus difficult to interpret. REVIGO is a Web server that summarizes long, unintelligible finding representative subset using simple clustering algorithm relies on semantic similarity measures. Furthermore, visualizes this non-redundant term set in multiple...
Automated annotation of protein function is challenging. As the number sequenced genomes rapidly grows, overwhelming majority products can only be annotated computationally. If computational predictions are to relied upon, it crucial that accuracy these methods high. Here we report results from first large-scale community-based critical assessment (CAFA) experiment. Fifty-four representing state art for prediction were evaluated on a target set 866 proteins 11 organisms. Two findings stand...
Detection of patient-zero can give new insights to the epidemiologists about nature first transmissions into a population. In this paper, we study statistical inference problem detecting source epidemics from snapshot spreading on an arbitrary network structure. By using exact analytic calculations and Monte Carlo estimators, demonstrate detectability limits for SIR model, which primarily depend process characteristics. Finally, applicability approach in case simulated sexually transmitted...
Bacteria and Archaea display a variety of phenotypic traits can adapt to diverse ecological niches. However, systematic annotation prokaryotic phenotypes is lacking. We have therefore developed ProTraits, resource containing ∼545 000 novel phenotype inferences, spanning 424 assigned 3046 bacterial archaeal species. These annotations were by computational pipeline that associates microbes with text-mining the scientific literature broader World Wide Web, while also being able define concepts...
Abstract Despite decades of intensive search for compounds that modulate the activity particular protein targets, a large proportion human kinome remains as yet undrugged. Effective approaches are therefore required to map massive space unexplored compound–kinase interactions novel and potent activities. Here, we carry out crowdsourced benchmarking predictive algorithms kinase inhibitor potencies across multiple families tested on unpublished bioactivity data. We find top-performing...
Codon usage bias in prokaryotic genomes is largely a consequence of background substitution patterns DNA, but highly expressed genes may show preference towards codons that enable more efficient and/or accurate translation. We introduce novel approach based on supervised machine learning detects effects translational selection genes, while controlling for local variation nucleotide represented as sequence composition intergenic DNA. A cornerstone our method Random Forest classifier...
The present paper demonstrates the antiproliferative ability and structure−activity relationships (SAR) of 14 crown aza-crown ether analogues on five tumor-cell types. most active compounds were di-tert-butyldicyclohexano-18-crown-6 (3), which exhibited cytotoxicity in submicromolar range, di-tert-butyldibenzo-18-crown-6 (5) (IC50 values ∼2 μM). Also, 3 5 induced marked influence cell cycle phase distributionstrong G1 arrest, followed by induction apoptosis. A computational SAR modeling...
P-glycoprotein (P-gp, MDR1) is a promiscuous drug efflux pump of substantial pharmacological importance. Taking advantage large-scale cytotoxicity screening data involving 60 cancer cell lines, we correlated the differential biological activities ∼13,000 compounds against cellular P-gp levels. We created large set 934 high-confidence substrates or nonsubstrates by enforcing agreement with an orthogonal criterion overexpressing ADR-RES cells. A support vector machine (SVM) was 86.7% accurate...
New microbial genomes are sequenced at a high pace, allowing insight into the genetics of not only cultured microbes, but wide range metagenomic collections such as human microbiome. To understand deluge genomic data we face, computational approaches for gene functional annotation invaluable. We introduce novel model that refines two established concepts: based on homology and phyletic profiling. The profiling-based includes both inferred orthologs paralogs—homologs separated by speciation...
Abstract Widespread use of herbicides results in the global increase weed resistance. The rotational according to their modes action (MoAs) and discovery novel phytotoxic molecules are two strategies used against Herein, Random Forest modeling was build predictive models establish comprehensive characterization structure–activity relationships underlying herbicide classifications MoAs selectivity. By combining with herbicide-likeness rules defined by selected molecular features (numbers...
The purpose of this study was the identification and quantification biochemical parameters over a 1-year cycle to provide detailed picture seasonal changes in plasma metabolites enzymes. Using novel methods machine learning techniques, authors created generated for first time comprehensible classification models exploring importance blood chemistry parameters, strength, mutual interactions or dependencies, reliability particular within groups.
Motivated by recent financial crises significant research efforts have been put into studying contagion effects and herding behaviour in markets. Much less has said about influence of news on We propose a novel measure collective the Web, News Cohesiveness Index (NCI), show that it can be used as systemic risk indicator. evaluate NCI documents from large Web sources daily basis October 2011 to July 2013 analyse interplay between markets financially related news. hypothesized strong cohesion...
Based on a set of subjects and collection attributes obtained from the Alzheimer's Disease Neuroimaging Initiative database, we used redescription mining to find interpretable rules revealing associations between those determinants that provide insights about disease (AD). We extended CLUS-RM algorithm constraint-based (CBRM) setting, which enables several modes targeted exploration specific, user-constrained associations. Redescription enabled finding specific constructs clinical biological...
Healthcare workplaces lack vital digital skills. In 2023, co-design sessions identified 20 essential learning objectives (LOs) aligned with EU policies.
Abstract Background Prokaryotic environmental adaptations occur at different levels within cells to ensure the preservation of genome integrity, proper protein folding and function as well membrane fluidity. Although specific composition structure cellular components suitable for variety extreme conditions has already been postulated, a systematic study describing such not yet performed. We therefore explored whether niche prokaryote could be deduced from sequence its proteome. Finally, we...
In this paper we introduce a statistical inference framework for estimating the contagion source from partially observed spreading process on an arbitrary network structure. The is based maximum likelihood estimation of partial epidemic realization and involves large scale simulation processes set potential locations. We present number different estimators that are used to determine conditional probabilities associated observing with particular location candidates. This also applicable...
Abstract Genes with similar roles in the cell cluster on chromosomes, thus benefiting from coordinated regulation. This allows gene function to be inferred by transferring annotations genomic neighbors, following guilt-by-association principle. We performed a systematic search for co-occurrence of >1000 functions neighborhoods across 1669 prokaryotic, 49 fungal and 80 metazoan genomes, revealing prevalent patterns that cannot explained clustering functionally genes. It is very common...
A comparative study of blood chemistry and histology was conducted on two groups mullets (Mugilidae) living under different conditions with feed sources. The aquaculture influenced mullet group (AIM), collected near fish farms the control (CM) caught in waters without any activities. Histological biochemical procedures were employed to liver histomorphology, plasma aspartate alanine aminotransferase (AST, ALT), triglyceride (TRIG), cholesterol (CHOL), glucose (GLU) total protein (TP) both...
We aim to demonstrate that a complex plant tissue protein mixture can be reliably "fingerprinted" by running conventional 1-D SDS-PAGE in bulk and analyzing gel banding patterns using machine learning methods. An unsupervised approach filter noise systemic biases (principal component analysis) was coupled state-of-the-art supervised methods for classification (support vector machines) attribute ranking (ReliefF) improve discrimination, visualization, recognition of important regions.