Pregled bibliografske jedinice broj: 1269049
QSAR bioactivity prediction in clinical metabolomics
QSAR bioactivity prediction in clinical metabolomics // 28. HSKIKI Book of Abstracts / Rogošić, Marko (ur.).
Zagreb: Hrvatsko društvo kemijskih inženjera i tehnologa (HDKI), 2023. str. 47-47 (predavanje, međunarodna recenzija, sažetak, znanstveni)
CROSBI ID: 1269049 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
QSAR bioactivity prediction in clinical metabolomics
Autori
Lovrić, Mario ; Widdowson, Michael ; Chawes, Bo ; Rasmussen, Morten
Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni
Izvornik
28. HSKIKI Book of Abstracts
/ Rogošić, Marko - Zagreb : Hrvatsko društvo kemijskih inženjera i tehnologa (HDKI), 2023, 47-47
Skup
28. Hrvatski skup kemičara i kemijskih inženjera
Mjesto i datum
Rovinj, Hrvatska, 28.03.2023. - 31.03.2023
Vrsta sudjelovanja
Predavanje
Vrsta recenzije
Međunarodna recenzija
Ključne riječi
QSAR ; bioactivity ; metabolomics ; machine learning
Sažetak
Maternal pregnancy dietary intake, nutrition, and environmental exposure in the early postnatal period of a newborn child are of importance for its health. To understand such relationships, we investigated the metabolomics in a mother-child cohort of n=602 mother child dyads. Amongst the metabolites, there were 79 quantified and annotated xenobiotics. Toxicological risk of the metabolites was evaluated by means of quantitative-structure activity relationships (QSARs). The QSARs were trained using Random forests tuned with Bayesian optimization on 203 molecular descriptors and 4096 Morgan fingerprints as features. The underlying training data were the Tox21 sets which consists of 12 bioactivity endpoints based on stress response and nuclear receptors from a total of n=8174 molecules. Metabolite structures were extracted using an automated pipeline for PubChem written in Python. The validated models were then applied to the metabolite data, which gave insight to which metabolites have a bioactive potential towards the selected endpoints. Results show an overlap in chemical spaces between the metabolites and the Tox21 data. Furthermore, there are many metabolites marked as bioactive, indicating their potential health risks if dysregulated in the body.
Izvorni jezik
Engleski
Znanstvena područja
Kemija, Interdisciplinarne prirodne znanosti, Računarstvo, Interdisciplinarne tehničke znanosti, Biotehnologija u biomedicini (prirodno područje, biomedicina i zdravstvo, biotehničko područje)
