Table 2 Summary of correlation and genomic prediction results for each trait

From: Robotized indoor phenotyping allows genomic prediction of adaptive traits in the field

 

Method indoora

Method fieldb

Experiment

#

Hyb

Observed genotypic value

Genomic prediction (genotypic means over experiments)

r (BLUEs correlation)c

rg (genetic correlation)d

Effe

#

Hyb

r (cross-validation)f

Acc (cross-validation)g

#

Hyb

r (external validation)h

Acc (external validation)i

Leaf appearance rate (LAR)

Time course

Time course

Indoor 2 VS Field 6

44

0.73 ± 0.07

0.69 ± 0.08

0.70 ± 0.08

302

0.58 ± 0.09

0.74 ± 0.06

50

0.53 ± 0.10

0.71 ± 0.08

Indoor 1 VS Field 3

21

0.57 ± 0.16

0.43 ± 0.19

0.45 ± 0.19

Field 5 VS Field 6

44

0.71 ± 0.08

0.65 ± 0.09

0.67 ± 0.09

Field 1 VS Field 3

26

0.49 ± 0.16

0.41 ± 0.17

0.41 ± 0.17

Vegetative phase duration

Irrelevant

Recorded

Field 5 VS Field 6

44

0.88 ± 0.04

0.77 ± 0.06

0.76 ± 0.07

302

0.84 ± 0.04

0.93 ± 0.02

60

0.71 ± 0.07

0.86 ± 0.04

Field 1 VS Field 3

53

0.69 ± 0.07

0.60 ± 0.09

0.62 ± 0.09

Field 2 VS Field 3

55

0.47 ± 0.11

0.40 ± 0.12

0.41 ± 0.12

Architecture (rhPAD / ALA)

3D imaging

UAV imaging

Indoor 2 VS Field 5

56

0.77 ± 0.06

0.66 ± 0.08

0.66 ± 0.08

302

0.65 ± 0.08

0.75 ± 0.06

20

0.42 ± 0.20

0.57 ± 0.16

Indoor 1 VS

Field 4

18

0.58 ± 0.17

0.45 ± 0.20

0.44 ± 0.21

Indoor 1 VS Field 2

18

0.60 ± 0.17

0.50 ± 0.19

0.49 ± 0.20

Field 4 VS Field 2

18

0.50 ± 0.19

0.41 ± 0.21

0.41 ± 0.21

Stomatal conductance (gsmax)

Model inversion

Not measurable at HTP

Indoor

–

–

–

–

302

0.56 ± 0.09

0.81 ± 0.05

–

–

–

Leaf expansion rate (LER)

Time course

Indoor

–

–

–

–

302

0.76 ± 0.06

0.92 ± 0.02

20

0.34 ± 0.21

0.46 ± 0.19

Leaf area index (LAI)

Modeling

UAV imaging

Indoor 2 VS Field 5_WW

51

0.64 ± 0.09

0.55 ± 0.10

0.53 ± 0.10

–

–

–

–

–

–

Indoor 2 VS Field 5_WD

51

0.44 ± 0.11

0.38 ± 0.12

0.37 ± 0.12

–

–

–

–

–

–

  1. a, bMethod for trait measurement indoor and in field, respectively. HTP, high-throughput phenotyping. cCorrelation between genotypic values (BLUEs) for each couple of experiments. dGenetic correlation between experiments assessed using a multivariate mixed model38, 39. eEfficiency of indirect selection, (case of an indirect phenotypic selection based on trait observed values in a given experiment, indoor or in a field), calculated as the accuracy of indirect selection divided by the square root of trait genomic heritability in the target field experiment52. fCorrelation between G-BLUP predicted values and measured values in a cross-validation scheme with diversity and genetic progress panels. gPrediction accuracy of genomic selection, calculated as in fdivided by the square root of trait genomic heritability,53. hExternal validation: Correlation between G-BLUP predicted values (with training on diversity and genetic progress panels) and measured values in recent hybrids panel. iPrediction accuracy, calculated as in hdivided by the square root of trait genomic heritability,53. Standard error (SE) estimates77 are shown after the ± symbol. For details, see Supplementary Tables 4 and 5.
Back to article page