Score test variable screening.

Area of Science:

• Genomics
• Biostatistics
• Computational Biology

Background:

• High-throughput data analysis requires efficient variable screening.
• Existing methods are often computationally intensive or theoretically complex.
• Genomic studies, like multiple myeloma, present high-dimensional challenges.

Purpose of the Study:

• To develop a computationally efficient and theoretically sound variable screening framework.
• To introduce a reproducible method for selecting the number of variables post-screening.
• To propose an iterative score test screening method linked to sparse regression.

Main Methods:

• A score test-based screening framework is established.
• A resampling procedure is proposed for variable selection.
• An iterative score test screening method is developed, related to sparse regression.

Main Results:

• The proposed score test screening framework is widely applicable and computationally efficient.
• Simulation studies demonstrate superior performance compared to existing methods across four regression models.
• The methods were successfully applied to multiple myeloma gene expression data.

Conclusions:

• Score test-based screening offers a robust and efficient approach for high-throughput data.
• The developed methods enhance reproducibility in variable selection.
• This framework is valuable for identifying key genes in complex diseases like multiple myeloma.