Control-Normalized Fisher Ratio Analysis of Comprehensive Two-Dimensional Gas Chromatography Time-of-Flight Mass Spectrometry Data for Enhanced Biomarker Discovery in a Metabolomic Study of Orthopedic Knee-Ligament Injury

Abstract

An innovative form of Fisher ratio (F-ratio) analysis (FRA) is developed for use with comprehensive two-dimensional gas chromatography/time-of-flight mass spectrometry (GC × GC−TOFMS) data and applied to the investigation of the changes in the metabolome in human plasma for patients with injury to their anterior cruciate ligament (ACL). Specifically, FRA provides a supervised discovery of metabolites that express a statistically significant variance in a two-sample class comparison: patients and healthy controls. The standard F-ratio utilizes the between-class variance relative to the pooled within-class variance. Because standard FRA is adversely impacted by metabolites expressed with a large within-class variance in the patient class, "control-normalized FRA" has been developed to provide complementary information, by normalizing the between-class variance to the variance of the control class only. Thirty plasma samples from patients who recently suffered from an ACL injury, along with matched controls, were subjected to GC × GC−TOFMS analysis. Following both standard and control-normalized FRA, the concentration ratio for the top 30 "hits" in each comparison was obtained and then t-tested for statistical significance. Twenty four out of 30 metabolites plus the therapeutic agent, naproxen (24/30), passed the t-test for the control-normalized FRA, which included 8/24 unique to control-normalized FRA and 16/24 in common with the standard FRA. Likewise, standard FRA provided 21/30 metabolites passing the t-test, with 5/21 undiscovered by control-normalized FRA. The complementary information obtained by both F-ratio analyses demonstrates the general utility of the new approach for a variety of applications.

A Meta-Analysis of Systemic Evaluation of Knee Ligament Injury or Intervention of Knee Proprioceptive Function Recovery

“…And the incidence of joint cruciate ligament injury is significantly higher in women than in men (2–4 times). In addition to women's exercise methods, lack of preparation for exercise training, and lack of proficiency in the initial training, it is mainly related to women's physiological characteristics [ 1 ]. For example, neuromuscular, joint ligament, periodic changes of hormones, female physical characteristics, and other factors are all related to women prone to knee cruciate ligament injury.…”

Analytical Determination of the Severity of Potato Taste Defect in Roasted East African Arabica Coffee

“…29 Application of a class-normalized Fratio calculation can discover these features that exhibit a high between-class variance relative to the within-class variance of one group. 29,30 This calculation can be potentially useful given the inherent biological variation present in foods such as coffee due to growing conditions, genetics, and post-harvest processing. 36−40 The C F-ratio, which uses ratios of betweenclass variance to the within-class variance of the clean samples, discovered 49 analytes of interest (Table S4).…”

“…F -ratio analysis can potentially be limited if a chemically meaningful analyte exhibits a large within-class variance or non-normal distribution . Application of a class-normalized F -ratio calculation can discover these features that exhibit a high between-class variance relative to the within-class variance of one group. , This calculation can be potentially useful given the inherent biological variation present in foods such as coffee due to growing conditions, genetics, and post-harvest processing. − The C F -ratio, which uses ratios of between-class variance to the within-class variance of the clean samples, discovered 49 analytes of interest (Table S4).…”

“…29 Additionally, two separate class-normalized F-ratio calculations were performed by dividing the between-class variance by the within-class variance of one group or the other group. 29,30 Herein, this calculation is referred to as either the clean-normalized ( C F-ratio) or strongnormalized ( S F-ratio) F-ratio depending upon the group used in the denominator. F-ratio analysis was performed for every data point above a signal-to-noise (S/N) threshold of 10 with at least six samples above the S/N threshold.…”

“…Due to the presence of two outlier samples in the strong PTD class (discussed later), F -ratio analysis was performed using all 13 clean samples and the remaining 13 strong PTD samples. The traditional F -ratio ( T F -ratio) calculation was calculated as the ratio of the between-class variance to the pooled within-class variance, referred to as the standard F -ratio calculation in previous work . Additionally, two separate class-normalized F -ratio calculations were performed by dividing the between-class variance by the within-class variance of one group or the other group. , Herein, this calculation is referred to as either the clean-normalized ( C F -ratio) or strong-normalized ( S F -ratio) F -ratio depending upon the group used in the denominator.…”