Skip to main content

Table 1 Simulated data, regression of results from logistic regression of pooled data on results from least squares regression of individual data

From: Estimating the effect of SNP genotype on quantitative traits from pooled DNA samples

Number of pools

3000

1024

512

256

128

64

32

16

8

4

2

Parameter

Statistic

Subset

With no genotyping error simulated

α

R2

All SNP

0.998

0.998

0.998

0.998

0.998

0.998

0.998

0.996

0.994

0.987

0.957

  

MLP < 5

0.999

0.999

0.999

0.999

0.999

0.998

0.997

0.992

0.983

0.955

0.849

 

Slope

All SNP

0.967

0.967

0.967

0.967

0.966

0.967

0.967

0.967

0.966

0.962

0.965

  

MLP < 5

0.997

0.998

0.998

0.997

0.998

0.996

0.995

0.989

0.979

0.980

0.993

MLP

R2

All SNP

0.999

0.999

0.999

0.999

0.999

0.999

0.999

0.998

0.997

0.993

0.978

  

MLP < 5

0.999

0.999

0.999

0.999

0.998

0.997

0.993

0.983

0.964

0.903

0.686

 

Slope

All SNP

0.923

0.923

0.922

0.922

0.922

0.921

0.918

0.909

0.886

0.819

0.622

  

MLP < 5

0.997

0.997

0.996

0.997

0.998

0.994

0.991

0.971

0.920

0.825

0.618

   

With genotyping error (SD = 0.05) simulated

α

R2

All SNP

-

0.975

0.929

0.895

0.861

0.808

0.705

0.597

0.388

0.110

0.024

  

MLP < 5

-

0.911

0.693

0.527

0.385

0.318

0.181

0.100

0.059

0.004

0.015

 

Slope

All SNP

-

0.929

0.900

0.903

0.915

0.894

0.908

0.911

0.947

0.829

0.873

  

MLP < 5

-

0.773

0.618

0.587

0.543

0.572

0.639

0.503

0.721

-0.370

-1.586

MLP

R2

All SNP

-

0.996

0.988

0.982

0.971

0.942

0.881

0.819

0.677

0.455

0.195

  

MLP < 5

-

0.909

0.765

0.680

0.577

0.334

0.139

0.076

0.053

0.039

0.016

 

Slope

All SNP

-

0.910

0.907

0.916

0.928

0.909

0.917

0.902

0.906

0.758

0.531

  

MLP < 5

-

0.924

0.882

0.871

0.831

0.823

0.757

0.670

0.953

1.424

1.112

  1. Simulated data, 3000 individuals, regression of results from logistic regression (genotype on phenotype, 3000 individuals or pooled data) on results from least squares regression (phenotype on genotype, 3000 individuals); parameters are the estimated effect (α) and minus the log p-value (MLP); the coefficient of determination (R2) and slope of the regression line are shown; results are presented for all SNP, or for SNP with MLP from individual data less than 5 (p-value > 1E-05), and with no genotyping error simulated, or with genotyping error (SD = 0.05) simulated