A Fisher scoring algorithm for the weighted regression method of QTL mapping

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Quantitative trait loci (QTL) mapping is crucial for understanding genetic contributions to complex traits.
• Existing methods like least squares (LS) and weighted LS have limitations in efficiency and parameter estimation.
• The estimating equation (EE) algorithm offers an implicit iterative approach for QTL analysis.

Purpose of the Study:

• To develop an explicit and programmable Fisher-scoring algorithm for the weighted LS method in QTL mapping.
• To enable automatic computation of the variance-covariance matrix for estimated QTL parameters.
• To facilitate the W-test statistic for assessing QTL significance.

Main Methods:

• Implementation of a Fisher-scoring algorithm for the weighted LS method.
• Development of a simplified method for computing the variance-covariance matrix under the expectation maximization (EM) algorithm.
• Comparison of the Fisher-scoring algorithm with the EM-based maximum likelihood (ML) method.

Main Results:

• The developed Fisher-scoring algorithm is explicit and easier to program than implicit EE algorithms.
• The method provides an approximate variance-covariance matrix as a byproduct, enabling W-test statistics.
• The new approach offers improved efficiency over standard LS and weighted LS methods.

Conclusions:

• The Fisher-scoring algorithm provides an efficient and practical alternative for weighted LS-based QTL mapping.
• The automatic computation of variance-covariance matrices simplifies significance testing for QTL.
• This advancement aids in the genetic dissection of complex traits through improved QTL analysis.