Matches in Ghent University Academic Bibliography for { <https://biblio.ugent.be/publication/01GV38DAQE41M2EVX64AGME6F3> ?p ?o. }
Showing items 1 to 21 of
21
with 100 items per page.
- 01GV38DAQE41M2EVX64AGME6F3 classification C3.
- 01GV38DAQE41M2EVX64AGME6F3 date "2022".
- 01GV38DAQE41M2EVX64AGME6F3 language "eng".
- 01GV38DAQE41M2EVX64AGME6F3 type conference.
- 01GV38DAQE41M2EVX64AGME6F3 subject "Medicine and Health Sciences".
- 01GV38DAQE41M2EVX64AGME6F3 subject "Veterinary Sciences".
- 01GV38DAQE41M2EVX64AGME6F3 presentedAt urn:uuid:56f8762c-8a2c-4843-a48e-34ff7c934d75.
- 01GV38DAQE41M2EVX64AGME6F3 abstract "The prevalence of food-related allergies is increasing worldwide, especially in industrialized countries, and is considered a major public health threat. This is particularly true at a young age, where one of the first allergies to develop is cow's milk allergy (CMA), an allergic reaction to cow's milk proteins (casein or beta-lactoglobulin) that is often associated with a higher risk of other atopic manifestations later in life. However, current diagnostic methods for food allergy lack either sensitivity (atopy patch tests) or specificity (serum IgE levels, skin prick tests), resulting in patients being under- or overtreated with unnecessary dietary restrictions. In search of more accurate diagnostic/prognostic markers, a lipidomics study was performed on stool and urine samples of children with confirmed IgE-mediated CMA (n=21), non-IgE-mediated CMA (n=19), IgE-mediated food allergy to other allergens (n=11), or healthy siblings (n=21). This resulted in the detection of e.g. 65913 metabolic features for stool across the 67 to 2300 Da range. To facilitate the selection of biologically relevant metabolites, the dataset was reduced to features with an FDR-corrected p-value (t-test) ≥ 0.1 and log (FC) ≥ 1, leaving 690 biomarker candidates. In addition, classification methods including (O)PLS-DA, Random Forest (RF), and Support Vector Machine (SVM) were tested to select the best approach for building predictive models. SVM method outperformed OPLS-DA and RF with a class prediction accuracy of 89% and a multivariate ROC AUC of 0.949 (CI 0.8-1) versus 76%, 0.858 (CI 0.67-1) and 75%, 0.803 (CI 0.57-0.99), respectively. To further characterize the retained molecules, their contribution to the multivariate statistics (e.g. variable importance in the projection, importance in the model) and the univariate ROC AUC values were assessed. After filtering for adduct and isotopic peaks, the top 100 components were selected for identification based on MS2 fragmentation spectra. Among the identified metabolites were several bile acids, including cholic acid, lithocholic acid, ursodeoxycholic acid, and hyodeoxycholic acid, as well as long-chain fatty acids, consistent with previous findings in the fecal metabolome of CMA patients and in our murine CMA model. Interestingly, the highest individual biomarker ROC AUC was 0.86 (CI 0.736-0.949), whereas a 7 biomarker combination drove the predictive accuracy to 96% with ROC AUC 0.997 (CI 0.942 - 1), emphasizing the need for multivariate assessment of the biomarker panel to account for potential interactions for optimal diagnostic performance. In summary, our study highlights the importance of multiple assessments of biomarker candidates using multi- and univariate statistical analyzes as well as wrapper-based machine learning methods to obtain the most biologically relevant molecules considering their interactions and presents a novel lipidomics biomarker panel to aid in the diagnosis of pediatric CMA.".
- 01GV38DAQE41M2EVX64AGME6F3 author 2D28CD88-F0EE-11E1-A9DE-61C894A0A6B4.
- 01GV38DAQE41M2EVX64AGME6F3 author CFD1E364-6721-11E6-9459-1DCEB4D1D7B1.
- 01GV38DAQE41M2EVX64AGME6F3 author F41CDCBE-F0ED-11E1-A9DE-61C894A0A6B4.
- 01GV38DAQE41M2EVX64AGME6F3 author F4500670-F0ED-11E1-A9DE-61C894A0A6B4.
- 01GV38DAQE41M2EVX64AGME6F3 author F77AFAE4-F0ED-11E1-A9DE-61C894A0A6B4.
- 01GV38DAQE41M2EVX64AGME6F3 author F815AD64-F0ED-11E1-A9DE-61C894A0A6B4.
- 01GV38DAQE41M2EVX64AGME6F3 dateCreated "2023-03-09T13:19:19Z".
- 01GV38DAQE41M2EVX64AGME6F3 dateModified "2024-10-29T08:53:58Z".
- 01GV38DAQE41M2EVX64AGME6F3 name "Novel lipidomic biomarker panel towards improving of diagnostic accuracy in pediatric cow's milk allergy".
- 01GV38DAQE41M2EVX64AGME6F3 sameAs LU-01GV38DAQE41M2EVX64AGME6F3.
- 01GV38DAQE41M2EVX64AGME6F3 sourceOrganization urn:uuid:19660bdd-ab07-4dc3-866c-854eb66d9370.
- 01GV38DAQE41M2EVX64AGME6F3 sourceOrganization urn:uuid:714020fd-edef-4225-8e7c-590e86c7ff5b.
- 01GV38DAQE41M2EVX64AGME6F3 type C3.