Table 1 Dataset characteristics used in model building, internal and external validation
From: Validating biomarkers and models for epigenetic inference of alcohol consumption from blood
Study | N | Age (years), mean (SD) | Men (%) | BMI mean (SD) | Alcohol gr/day, Median (min, max) | Non-drinkers (%) | Light drinkers (%) | At-risk drinkers (%) | Heavy drinkers (%) |
|---|---|---|---|---|---|---|---|---|---|
Model building and internal validation dataset | |||||||||
RS-II-3/III-2 | 611 | 67 (6) | 275 (45) | 27.8 (4) | 8.6 (1, 57) | 0 (0) | 545 (89) | 52 (9) | 14 (2) |
CODAM | 159 | 66 (7) | 86 (54) | 28.9 (4) | 7.9 (0, 72) | 12 (8) | 117 (74) | 23 (14) | 7 (4) |
NTR | 617 | 39 (14) | 188 (31) | 24.6 (4) | 5.1 (0, 69) | 195 (32) | 348 (56) | 44 (7) | 30 (5) |
LLS | 491 | 58 (6) | 231 (47) | 25.3 (3) | 13.0 (0, 90) | 36 (7) | 309 (63) | 98 (20) | 49 (10) |
PAN | 164 | 62 (9) | 100 (61) | 26.0 (4) | 11.0 (0, 77) | 1 (1) | 127 (77) | 20 (12) | 16 (10) |
KORA F4 | 841 | 61 (9) | 415 (49) | 28.0 (5) | 7.6 (0, 150) | 251 (30) | 354 (42) | 133 (16) | 103 (12) |
Total dataset | 2883 | 57 (14) | 1295 (45) | 26.7 (4) | 8.0 (0, 150) | 495 (17) | 1800 (62) | 370 (13) | 218 (8) |
External validation datasets | |||||||||
SHIP-Trend | 433 | 51 (14) | 205 (47) | 27.2 (4.1) | 3.6 (0, 82) | 47 (11) | 346 (80) | 28 (6) | 12 (3) |
TwinsUK | 713 | 58 (10) | 0 (0) | 26.7 (5) | 2.3 (0, 101) | 187 (26) | 423 (59) | 67 (9) | 36 (5) |
TwinsUK2 | 442 | 59 (9) | 0 (0) | 26.6 (5) | 5.3 (0, 94) | 36 (8) | 311 (70) | 46 (10) | 49 (11) |
RS-III-1 | 648 | 59.6 (8) | 298 (46) | 27.7 (5) | 6.4 (0, 57) | 64 (10) | 495 (76) | 79 (12) | 10 (2) |