Research of inflammatory bowel disease (IBD) has identified numerous molecular players involved in the disease development. Even so, the understanding of IBD is incomplete, while disease treatment is still far from the precision medicine. Reliable diagnostic and prognostic biomarkers in IBD are limited which may reduce efficient therapeutic outcomes. High-throughput technologies and artificial intelligence emerged as powerful tools in search of unrevealed molecular patterns that could give important insights into IBD pathogenesis and help to address unmet clinical needs. Machine learning, a subtype of artificial intelligence, uses complex mathematical algorithms to learn from existing data in order to predict future outcomes. The scientific community has been increasingly employing machine learning for the prediction of IBD outcomes from comprehensive patient data-clinical records, genomic, transcriptomic, proteomic, metagenomic, and other IBD relevant omics data. This review aims to present fundamental principles behind machine learning modeling and its current application in IBD research with the focus on studies that explored genomic and transcriptomic data. We described different strategies used for dealing with omics data and outlined the best-performing methods. Before being translated into clinical settings, the developed machine learning models should be tested in independent prospective studies as well as randomized controlled trials.

Celiac disease (CD) is a chronic inflammatory disease in the small intestine triggered by gluten uptake that occurs in genetically susceptible individuals. HLA-DQ2 protein encoded by HLA-DQA1*05 and DQB1*02 alleles is found in 90-95% of CD patients. All of the remaining patients carry HLA-DQ8 protein encoded by HLA-DQA1*03 and DQB1*03:02 alleles. Specific HLA-DQ genotypes define different risk for CD incidence. Presence of susceptible HLA-DQ genotypes does not predict certain disease development, but their absence makes CD very unlikely, close to 100%. Here we presented for the first time the distribution of HLA-DQ genotypes in the group of pediatric celiac patients from the University Children’s Hospital, Belgrade, Serbia and estimated risk for CD development that these genotypes confer. Seventy three celiac disease patients and 62 healthy individuals underwent genotyping for DQA1, DQB alleles and DRB1 allele. 94.5% of patients carried alleles that encode DQ2 protein variant and 2.7% carried alleles that encode DQ8 protein variant. Two patients carried single DQB1*02 allele. No patients were negative for all the alleles predisposing to CD. The highest HLA-DQ genotype risk for CD development was found in group of patients homozygous for DQ2.5 haplotype, followed by the group of heterozygous carriers of DQ2.5 haplotype in combination with DQB1*02 allele within the other haplotype. The lowest risk was observed in carriers of a single copy of DQB1*02 or DQA1*05 allele or other non-predisposing alleles. HLA genotyping, more informative than serological testing commonly used, proved to be a useful diagnostic tool for excluding CD development.

Leukemia is a heterogenous group of hematological malignancies categorized in four main types (acute myeloid leukemia (AML), acute lymphoblastic leukemia (ALL), chronic myeloid leukemia (CML) and chronic lymphocytic leukemia (CLL). Several cytogenetic and molecular markers have become a part of routine analysis for leukemia patients. These markers have been used in diagnosis, risk-stratification and targeted therapy application. Recent studies have indicated that numerous regulatory RNAs, such as long non-coding RNAs (lncRNAs), have a role in tumor initiation and progression. When it comes to leukemia, data for lncRNA involvement in its etiology, progression, diagnosis, treatment and prognosis is limited. The aim of this review is to summarize research data on lncRNAs in different types of leukemia, on their expression pattern, their role in leukemic transformation and disease progression. The usefulness of this information in the clinical setting, i.e., for diagnostic and prognostic purposes, will be emphasized. Finally, how particular lncRNAs could be used as potential targets for the application of targeted therapy will be considered.

Phenylketonuria (PKU) is caused by mutations in the gene encoding phenylalanine hydroxylase (PAH) enzyme. Here, we report the updated spectrum of PAH mutations in 61 Serbian PKU patients. By using both DGGE/DNA sequencing and PCR-RFLP, we identified 26 disease-causing mutations (detection rate 99%). The most frequent ones were p.L48S (31%), p.R408W (16.4%), p.P281L (6%), p.E390G (5.2%), and p.I306V (5.2%). Homozygosity value indicated high heterogeneity of Serbian population. To overcome possible pitfalls of patients’ phenotypic classification, we used two parameters: pretreatment/maximal phenylalanine blood concentration and Phe tolerance. The two phenotypes did not match only for patients with p.L48S. Therefore, we used Mann-Whitney statistical test to compare pretreatment/maximal blood Phe concentration and Phe tolerance detected in patients with p.[L48S];[null] and p.[missense];[null] genotypes. For patients with p.L48S, our results implied that Phe tolerance is a better parameter for phenotypic classification. Also, Fisher’s exact test was used to compare p.L48S effect on phenotype of homozygous and functionally hemizygous patients. Our findings showed that effect of p.L48S was altered in functional hemizygotes. Moreover, phenotypic inconsistency found in homozygotes suggested that interallelic complementation and/or additional factors play a role in genotype-phenotype correlation. Since BH4-supplementation therapy is not available in Serbia, we made the first estimation of its potential benefit based on patients’ genotypes. In the analyzed cohort, the total frequency of BH4-responsive mutations was 52.6%. Furthermore, we found a significant number of genotypes (26.2% BH4-responsive and 51% probably BH4-responsive) that may respond to BH4 therapy. This led us to a conclusion that BH4-supplementation therapy could bring benefit to Serbian PKU patients.

Aims: 6-mercaptopurine influences in vitro TPMT gene expression in a TPMT promoter variable number of tandem repeats (VNTR)-dependent manner. We studied TPMT expression following 6-mercaptopurine a...