Bayesian Robust Linear Model With Mahalanobis (Brlmm) Distance Classifier Algorithm - Newport iServer MicroServer iTHX-M Operator's Manual

Bayesian Robust Linear Model with
Mahalanobis (BRLMM) Distance Classifier
Algorithm
The BRLMM (Bayesian Robust Linear Model with Mahalanobis)
distance classifier algorithm is an extension of Nusrat Rabbee and
Terry Speed's RLMM (Robust Linear Model with Mahalanobis)
distance classifier algorithm [1]. BRLMM provides a significant
improvement over the DM (Dynamic Model) algorithm in two
important areas. It improves overall performance (call rates and
accuracy) and balances performance on homozygous and heterozygous
genotypes.
This appendix describes genotype calling with BRLMM:
• Normalization and Allele Summarization
• Clustering Space Transformation
• Calling Genotypes
• Estimating Cluster Centers and Variances
• Special Cases
• Results
• Discussion
355