Identification of Reliable Components in Multivariate Curve Resolution-Alternating Least Squares (MCR-ALS): a Data-Driven Approach across Metabolic Processes.

There is an increasing need to use multivariate statistical methods for understanding biological functions, identifying the mechanisms of diseases, and exploring biomarkers. In addition to classical analyses such as hierarchical cluster analysis, principal component analysis, and partial least squares discriminant analysis, various multivariate strategies, including independent component analysis, non-negative matrix factorization, and multivariate curve resolution, have recently been proposed. However, determining the number of components is problematic. Despite the proposal of several different methods, no satisfactory approach has yet been reported. To resolve this problem, we implemented a new idea: classifying a component as "reliable" or "unreliable" based on the reproducibility of its appearance, regardless of the number of components in the calculation. Using the clustering method for classification, we applied this idea to multivariate curve resolution-alternating least squares (MCR-ALS). Comparisons between conventional and modified methods applied to proton nuclear magnetic resonance ((1)H-NMR) spectral datasets derived from known standard mixtures and biological mixtures (urine and feces of mice) revealed that more plausible results are obtained by the modified method. In particular, clusters containing little information were detected with reliability. This strategy, named "cluster-aided MCR-ALS," will facilitate the attainment of more reliable results in the metabolomics datasets.

Implementation of the modified-SHIRPA protocol for screening of dominant phenotypes in a large-scale ENU mutagenesis program.

SHIRPA is a three-stage protocol for the comprehensive assessment of primarily mouse behavior. The first stage consists of high-throughput phenotyping of 33 behavioral observations and 7 metabolic or disease observations. We modified this part of the protocol by integrating new morphologic observations into the initial phenotype assay of behavior and dysmorphology. Behavioral observations assessed by this protocol, now referred to as the "modified-SHIRPA," are compatible with the original "SHIRPA" protocol. Using modified-SHIRPA, we screened dominant phenotypes of more than 10,000 G(1) progeny generated by crossing DBA/2J females with ENU-treated C57BL/6J males. To date, we have obtained 136 hereditary-confirmed mutants that exhibit behavioral and morphologic defects. Some independent mutant lines exhibited similar phenotypes, suggesting that they may represent alleles of the same gene or mutations in the same genetic pathway. They could hold great potential for the unraveling of the molecular mechanisms of certain phenotypes.

A series of maturity onset diabetes of the young, type 2 (MODY2) mouse models generated by a large-scale ENU mutagenesis program.

Mutant mouse models are indispensable tools for clarifying the functions of genes and for elucidating the underlying pathogenic mechanisms of human diseases. Currently, several large-scale mutagenesis projects that employ the chemical mutagen N-ethyl-N-nitrosourea (ENU) are underway worldwide. One specific aim of our ENU mutagenesis project is to generate diabetic mouse models. We screened 9375 animals for dominant traits using a clinical biochemical test and thereby identified 11 mutations in the glucokinase (Gk) gene that were associated with hyperglycemia. GK is a key regulator of insulin secretion in the pancreatic beta-cell. Approximately 190 heterozygous mutations in the human GK gene have been reported to cause maturity onset diabetes of the young, type 2 (MODY2). In addition, five mutations have been reported to cause permanent neonatal diabetes mellitus (PNDM) when present on both alleles. The mutations in our 11 hyperglycemic mutants are located at different positions in Gk. Four have also been found in human MODY2 patients, and another mutant bears its mutation at the same location that is mutated in a PNDM patient. Thus, ENU mutagenesis is effective for developing mouse models for various human genetic diseases, including diabetes mellitus. Some of our Gk mutant lines displayed impaired glucose-responsive insulin secretion and the mutations had different effects on Gk mRNA levels and/or the stability of the GK protein. This collection of Gk mutants will be valuable for understanding GK gene function, for dissecting the function of the enzyme and as models of human MODY2 and PNDM.