Predicting Gestational Age Using Whole-blood RNAseq

Bioinformatics

Each year, preterm birth affects 15 million babies, and is the leading cause of infant mortality. To prevent preterm birth, clinicians need to measure the amount of time the baby is in the womb, otherwise known as gestational age (GA).

As of right now, methods to measure gestational age such as ultrasounds and monitoring periods exist, but can be expensive and inaccurate.

Objectives

Bioinformatics allows us to determine gestational age in a way that is less invasive, and less costly by analyzing maternal whole-blood transcriptomics (study of RNA transcripts). We are using transcriptomics to determine how active a gene is based on the number of mRNA fragments in a sample that map to that specific gene.

Transcriptomics

Our Process

Using RStudio, our project consisted of these main stages:

1) Data Analysis

loading our datasets, and analyzing the correlations between our variables

2) Feature Extraction

reducing the amount of features and selecting the most important features

3) Machine Learning Models

using training and test data to predict unknown gestational ages

Our project was based on the Preterm Birth: Transcriptomics DREAM Challenge. Our results in this contest can be found here.

Page updated

Google Sites

Report abuse