Table 1 Summary of the datasets used in this study

From: LEOPARD: missing view completion for multi-timepoint omics data via representation disentanglement and temporal knowledge transfer

 

MGH COVID proteomics

KORA metabolomics

KORA multi-omics

Extended KORA metabolomics

Omics type

proteomics

metabolomics

metabolomics; proteomics

metabolomics

View

\({{\rm{v}}}1\)

panel: Explore 384 cardiometabolic

biochemical class*: GPL

omics: metabolomics

biochemical class*: GPL

\({{\rm{v}}}2\)

panel: Explore 384 inflammation

biochemical class*: AC, SL, AA, MS

omics: proteomics

biochemical class*: AC, SL, AA, MS

Variable number

Explore 384 cardiometabolic: 322,

Explore 384 inflammation: 295

GPL: 70,

AC: 15,

SL: 11,

AA: 9,

MS: 1

metabolites: 104,

proteins: 66

GPL: 69,

AC: 14,

SL: 11,

AA: 9,

MS: 1

Timepoint

\({{\rm{t}}}1\)

D0

F4

S4

S4

\({{\rm{t}}}2\)

D3

FF4

F4

F4

\({{\rm{t}}}3\)

   

FF4

Time span

3 days

7 years

7 years

14 years in total

Sample number

Total (100%)

218

2085

1062

614

Training (64%)

140

1335

680

393

Validation (16%)

35

333

170

98

Test (20%)

43

417

212

123

Masked data for completion

inflammation proteins from D3

non-GPL metabolites from FF4

proteins from F4

completion 1: Non-GPL metabolites from S4;completion 2: GPL metabolites from FF4

  1. *The targeted metabolites from KORA are classified into five analyte classes, namely glycerophospholipid (GPL), acylcarnitine (AC), sphingolipid (SL), amino acid (AA), and monosaccharide (MS). One view group comprises metabolites from the glycerophospholipid class, while the other view group comprises metabolites from the other classes.