$\widehat{\mu}$

gooddown

goodup

poordown

poorup


HLAF

0.31

0.65

0.29

0.40

IGLC2

0.56

0.98

0.46

0.68

LY9

0.29

0.58

0.52

1.12

TNFRSF17

0.41

0.97

0.58

0.59

SPP1

0.01

0.38

0.47

0.57

XCL2

0.36

0.67

0.41

0.58

C1QA

0.39

0.79

0.40

0.57

$\widehat{\Sigma}$

0.74

0.74

0.58

0.58

$\widehat{\pi}$ ∝ I

0.31

0.28

0.32

0.09

 Estimated mean expression profiles $\widehat{\mu}$, covariance matrices $\widehat{\Sigma}$ and weights $\widehat{\pi}$ for the four subgroups, as estimated from the training set. Note that the optimal covariance matrices were all proportional to the identity matrix $\widehat{\Sigma}$ ∝ I and are thus summarised by a single value, the variance of expression of the corresponding cluster. C1QA, complement component 1, q subcomponent, A chain; HLAF, major histocompatibility complex, class I, F; IGLC2, immunoglobulin lambda constant 2; LY9, lymphocyte antigen 9; TNFRSF17, tumour necrosis factor receptor superfamily member 17; SPP1, secreted phosphoprotein 1 (osteopontin); XCL2, chemokine (C motif) ligand 2.