Disease's type
GDM
Experimental grouping
GDM(n=1942),Normal Glucose Tolerance(n=3143)
GPT's summary
This study explores the use of machine learning and low-coverage genome-wide sequencing of cell-free DNA (cfDNA) from maternal plasma to predict gestational diabetes mellitus (GDM) in early pregnancy. Using a dataset of 5085 pregnant women, including 1942 diagnosed with GDM, the researchers developed a deep learning model with an attention mechanism that achieved high accuracy in identifying GDM. The model also revealed several genes, such as DEFA1, DEFA3, and DEFB1, associated with GDM, suggesting potential biomarkers for early detection and management of the condition.
RF's name
Defensin Alpha 3
Sample's type
Plasma
Gestational weeks
12th to 18th gestational weeks
Experiemental methods
Whole Genome Sequencing
Machine learning algorithms
A deep neural network with an attention architecture was used for binary classification
Title
Identify gestational diabetes mellitus by deep learning model from cell-free DNA at the early gestation stage
Evidence's type
Risk factor
Year
2024
Journal
Briefings in Bioinformatics
PMID