Table 1 Evaluation of the fitness of the discriminant models and the relative importance of canonical functions within each model.
From: Multivariate indicators of disease severity in COVID-19
Model 1 | ||||||||
|---|---|---|---|---|---|---|---|---|
Suitability of data for discriminant analysis | ||||||||
Pooled within-groups matrices (correlation between predictor variables) | Box’s M test p-value (cutoff > 0.001) | |||||||
Pearson correlations | NK cells | ActSMB | T cells | ActNeut | ImGran | 5.7528 × 10−14 | ||
NK cells | 1 | |||||||
ActSMB | − 0.192 | 1 | ||||||
T cells | − 0.080 | 0.144 | 1 | |||||
ActNeut | 0.055 | − 0.534 | 0.117 | 1 | ||||
ImGran | 0.377 | − 0.339 | 0.120 | 0.184 | 1 | |||
Model fitness | ||||||||
|---|---|---|---|---|---|---|---|---|
Step | Wilks’ λ | p-value | ||||||
1 | 0.351 | 1.1055 × 10−8 | ||||||
2 | 0.189 | 1.0226 × 10−11 | ||||||
3 | 0.125 | 3.3432 × 10−13 | ||||||
4 | 0.083 | 1.261 × 10−14 | ||||||
5 | 0.065 | 6.6843 × 10−15 | ||||||
Model | 0.065 | 4.8677 × 10−15 | ||||||
Discriminant function discriminatory power | ||||||||
|---|---|---|---|---|---|---|---|---|
Eigenvalue | % Variance | Canonical correlation | ||||||
First discriminant function | 4.155 | 67.7 | 0.898 | |||||
Second discriminant function | 1.985 | 32.3 | 0.815 | |||||
Model 2 | ||||||||
|---|---|---|---|---|---|---|---|---|
Suitability of data for discriminant analysis | ||||||||
Pooled within-groups matrices (correlation between predictor variables) | Box’s M test p-value (cutoff > 0.001) | |||||||
Pearson correlations | MAIT | En38NK | ActNB | En27NK | Temra | 9.8324 × 10−22 | ||
MAIT | 1 | |||||||
En38NK | 0.147 | 1 | ||||||
ActNB | − 0.226 | − 0.254 | 1 | |||||
En27NK | 0.192 | − 0.202 | 0.297 | 1 | ||||
Temra | − 0.093 | 0.068 | − 0.039 | − 0.343 | 1 | |||
Model fitness | ||||||||
|---|---|---|---|---|---|---|---|---|
Step | Wilks’ λ | p-value | ||||||
1 | 0.459 | 1.4335 × 10−7 | ||||||
2 | 0.347 | 8.8178 × 10−9 | ||||||
3 | 0.224 | 3.6834 × 10−11 | ||||||
4 | 0.191 | 1.9134 × 10−11 | ||||||
5 | 0.166 | 1.3565 × 10−11 | ||||||
Model | 0.166 | 1.101 × 10−11 | ||||||
Discriminant function discriminatory power | ||||||||
|---|---|---|---|---|---|---|---|---|
Eigenvalue | % Variance | Canonical correlation | ||||||
First discriminant function | 5.033 | 100.0 | 0.913 | |||||
Model 3 | ||||||||
|---|---|---|---|---|---|---|---|---|
Suitability of data for discriminant analysis | ||||||||
Pooled within-groups matrices (correlation between predictor variables) | Box’s M Test p-value (cutoff > 0.001) | |||||||
Pearson correlations | Monocyte HLA-DR MFI | ActSMB | NK cells | CXCR5+CD8+ MAIT | cDC HLA-DR MFI | CD45RA+ effector memory CD8+ T cells | Neutrophils | N/A |
Monocyte HLA-DR MFI | 1 | |||||||
ActSMB | − 0.501 | 1 | ||||||
NK cells | 0.231 | − 0.333 | 1 | |||||
CXCR5+CD8+ MAIT | − 0.094 | − 0.252 | − 0.224 | 1 | ||||
cDC HLA-DR MFI | 0.097 | − 0.165 | − 0.158 | − 0.227 | 1 | |||
CD45RA+ effector memory CD8+ T cells | − 0.053 | − 0.022 | 0.207 | 0.374 | 0.156 | 1 | ||
Neutrophils | − 0.163 | − 0.007 | − 0.228 | 0.251 | 0.012 | − 0.397 | 1 | |
Model fitness | ||||||||
|---|---|---|---|---|---|---|---|---|
Step | Wilks’ λ | p-value | ||||||
1 | 0.413 | 3 × 10−6 | ||||||
2 | 0.196 | 3.1293 × 10−9 | ||||||
3 | 0.0.142 | 6.324 × 10−10 | ||||||
4 | 0.113 | 4.3273 × 10−10 | ||||||
5 | 0.095 | 4.9018 × 10−10 | ||||||
6 | 0.072 | 2.1319 × 10−10 | ||||||
7 | 0.058 | 2.0125 × 10−11 | ||||||
Model | 0.058 | 8.7331 × 10−11 | ||||||
Discriminant function discriminatory power | ||||||||
|---|---|---|---|---|---|---|---|---|
Eigenvalue | % Variance | Canonical correlation | ||||||
First discriminant function | 16.219 | 100.0 | 0.971 | |||||