Pregnancy Status Prediction From Cell-Free Gene Expression Profiles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting premature delivery in pregnant women, such as fetal fibronectin molecule detection and fetal cfDNA concentration, suffer from high false-positives and insufficient correlation, lacking an effective diagnostic method.
Innovation Solution
A method and system for constructing a prediction model using differentially expressed gene information of cell-free nucleic acids in peripheral blood, obtained via one-time blood sampling, to predict pregnancy status, including premature delivery, preeclampsia, and other complications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If fetal fibronectin molecule detection is used for predicting premature delivery, then prediction capability is provided, but false-positive rate increases significantly
Solution Approach 1:
The patent segments the prediction task by dividing it into multiple independent prediction models, each targeting specific pregnancy complications (premature delivery, preeclampsia, fetal growth restriction). Each model is trained on specific differentially expressed genes relevant to its target condition, allowing specialized accurate prediction for each condition rather than a single general model that increases false positives.
Solution Approach 2:
The patent applies local quality by selecting differentially expressed genes specific to each pregnancy complication type. For example, certain genes are selected for premature delivery prediction while other genes are selected for preeclampsia prediction. This localized gene selection improves prediction accuracy for each specific condition while reducing false positives compared to universal markers like fetal fibronectin.
2Device complexity
If fetal cfDNA concentration is used as a single marker for predicting premature delivery, then prediction approach is simplified, but correlation insufficiency occurs
Solution Approach 1:
The patent uses a composite approach by combining multiple differentially expressed genes into integrated prediction models for each pregnancy complication. Instead of relying on a single marker like fetal cfDNA concentration, the patent composite multiple gene expression profiles (e.g., combining several genes related to inflammation, placental function, and immune response) to create robust prediction models that achieve high correlation and accuracy.
3Measurement precision
If multiple blood sampling is performed for accurate pregnancy status detection, then diagnostic accuracy improves, but patient burden and time consumption increase
Solution Approach 1:
The patent applies preliminary action by performing comprehensive whole-genome sequencing and identifying all differentially expressed genes during the initial blood sampling. This preliminary comprehensive analysis allows the system to predict multiple pregnancy complications simultaneously from a single sample, eliminating the need for repeated sampling while maintaining high diagnostic accuracy through the use of multiple gene markers.
Data Source
AI summary
Provided is a method for constructing a prediction model used for predicting a pregnancy status of a pregnant woman. The method includes: step 1 of constructing a training set and an optional test set, the training set and the optional test set each consisting of a plurality of pregnant woman samples, the pregnant woman samples each having a known pregnancy status; step 2 of determining, for each pregnant woman sample in the training set, a predetermined parameter of the pregnant woman sample, the predetermined parameter including differentially expressed gene information of the pregnant woman sample, the differentially expressed gene information being obtained through calculation based on sequencing information of fetal cell-free nucleic acids in peripheral blood of the pregnant woman sample; and step 3 of constructing the prediction model based on the known pregnancy status and the predetermined parameter.