An empirical Bayes approach for constructing confidence intervals for clonality and entropy

Area of Science:

• Immunology
• Bioinformatics
• Statistical Genetics

Background:

• Quantifying human immune responses to environmental factors is crucial.
• Analyzing V(D)J gene rearrangements in blood samples provides insights into immune diversity.
• Existing methods lack robust confidence intervals for immune diversity parameters.

Purpose of the Study:

• To develop a statistical method for constructing confidence intervals for immune diversity parameters.
• To address the challenge of inferring population diversity from large-scale V(D)J sequencing data.
• To provide a robust framework for quantifying uncertainty in immune repertoire analysis.

Main Methods:

• Utilized Polymerase Chain Reaction (PCR) to amplify V(D)J sequences from blood samples.
• Employed an Empirical Bayes method coupled with resampling-based calibration.
• Developed a statistical inference approach for high-dimensional multinomial-type data.

Main Results:

• Proposed a novel, robust confidence interval construction for population diversity parameters.
• Demonstrated the method's efficacy through extensive numerical simulations.
• Validated the approach using real-world blood sample data.

Conclusions:

• The proposed Empirical Bayes and resampling method effectively constructs robust confidence intervals for immune diversity.
• This approach enhances the statistical rigor of analyzing V(D)J sequence data.
• The findings contribute to a better understanding of human immune responses and repertoire diversity.