nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

17240_en-US

liliconcom

Haplotype Regression

I am running the haplotype regression function and am unclear exactly how to interpret the output files. What I am most interested in is determining the trait mean for each haplotype. In the SAS output window I am given estimate values for each haplotype. These values are all negative for all haplotypes in the overall test while some are negative for some haplotypes and positive for others in the individual tests. Any help would be appreciated. I attached the output below:

----------------------------------------------------------------------------------------------------

Overall Test 16:33 Monday, February 29, 2016 20

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Number of Observations Read 132

Number of Observations Used 132

----------------------------------------------------------------------------------------------------

Overall Test 16:33 Monday, February 29, 2016 21

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Dependent Variable: PSS_2013

Sum of

Source DF Squares Mean Square F Value Pr > F

Model 7 0.57061574 0.08151653 5.13 <.0001

Error 124 1.97091759 0.01589450

Corrected Total 131 2.54153333

R-Square Coeff Var Root MSE PSS_2013 Mean

0.224516 59.09690 0.126073 0.213333

Source DF Type I SS Mean Square F Value Pr > F

H1 1 0.04265381 0.04265381 2.68 0.1039

H2 1 0.01099720 0.01099720 0.69 0.4071

H3 1 0.15427488 0.15427488 9.71 0.0023

H4 1 0.15181720 0.15181720 9.55 0.0025

H5 1 0.13429041 0.13429041 8.45 0.0043

H6 1 0.07494326 0.07494326 4.72 0.0318

H7 1 0.00163899 0.00163899 0.10 0.7487

Source DF Type III SS Mean Square F Value Pr > F

H1 1 0.00756075 0.00756075 0.48 0.4917

H2 1 0.03763456 0.03763456 2.37 0.1264

H3 1 0.29761793 0.29761793 18.72 <.0001

H4 1 0.24398325 0.24398325 15.35 0.0001

H5 1 0.17124082 0.17124082 10.77 0.0013

H6 1 0.07487811 0.07487811 4.71 0.0319

H7 1 0.00163899 0.00163899 0.10 0.7487

Contrast DF Contrast SS Mean Square F Value Pr > F

Overall 7 0.57061574 0.08151653 5.13 <.0001

Standard

Parameter Estimate Error t Value Pr > |t|

Intercept 0.2890451760 0.02424577 11.92 <.0001

H1 -.0302601288 0.04387445 -0.69 0.4917

H2 -.0674181711 0.04381340 -1.54 0.1264

H3 -.1993103565 0.04605998 -4.33 <.0001

H4 -.2019650826 0.05154893 -3.92 0.0001

H5 -.1608493234 0.04900484 -3.28 0.0013

H6 -.1115559265 0.05139710 -2.17 0.0319

----------------------------------------------------------------------------------------------------

Overall Test 16:33 Monday, February 29, 2016 22

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Dependent Variable: PSS_2013

Standard

Parameter Estimate Error t Value Pr > |t|

H7 -.0151106871 0.04705650 -0.32 0.7487

----------------------------------------------------------------------------------------------------

Test for Haplotype 1 16:33 Monday, February 29, 2016 23

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Number of Observations Read 132

Number of Observations Used 132

----------------------------------------------------------------------------------------------------

Test for Haplotype 1 16:33 Monday, February 29, 2016 24

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Dependent Variable: PSS_2013

Sum of

Source DF Squares Mean Square F Value Pr > F

Model 1 0.04265381 0.04265381 2.22 0.1387

Error 130 2.49887952 0.01922215

Corrected Total 131 2.54153333

R-Square Coeff Var Root MSE PSS_2013 Mean

0.016783 64.98936 0.138644 0.213333

Source DF Type I SS Mean Square F Value Pr > F

H1 1 0.04265381 0.04265381 2.22 0.1387

Source DF Type III SS Mean Square F Value Pr > F

H1 1 0.04265381 0.04265381 2.22 0.1387

Contrast DF Contrast SS Mean Square F Value Pr > F

H1 1 0.04265381 0.04265381 2.22 0.1387

Standard

Parameter Estimate Error t Value Pr > |t|

Intercept 0.2058495133 0.01307144 15.75 <.0001

H1 0.0604036959 0.04054951 1.49 0.1387

----------------------------------------------------------------------------------------------------

Test for Haplotype 2 16:33 Monday, February 29, 2016 25

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Number of Observations Read 132

Number of Observations Used 132

----------------------------------------------------------------------------------------------------

Test for Haplotype 2 16:33 Monday, February 29, 2016 26

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Dependent Variable: PSS_2013

Sum of

Source DF Squares Mean Square F Value Pr > F

Model 1 0.00464098 0.00464098 0.24 0.6266

Error 130 2.53689236 0.01951456

Corrected Total 131 2.54153333

R-Square Coeff Var Root MSE PSS_2013 Mean

0.001826 65.48180 0.139695 0.213333

Source DF Type I SS Mean Square F Value Pr > F

H2 1 0.00464098 0.00464098 0.24 0.6266

Source DF Type III SS Mean Square F Value Pr > F

H2 1 0.00464098 0.00464098 0.24 0.6266

Contrast DF Contrast SS Mean Square F Value Pr > F

H2 1 0.00464098 0.00464098 0.24 0.6266

Standard

Parameter Estimate Error t Value Pr > |t|

Intercept 0.2108940059 0.01314754 16.04 <.0001

H2 0.0199012940 0.04080901 0.49 0.6266

----------------------------------------------------------------------------------------------------

Test for Haplotype 3 16:33 Monday, February 29, 2016 27

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Number of Observations Read 132

Number of Observations Used 132

----------------------------------------------------------------------------------------------------

Test for Haplotype 3 16:33 Monday, February 29, 2016 28

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Dependent Variable: PSS_2013

Sum of

Source DF Squares Mean Square F Value Pr > F

Model 1 0.18816586 0.18816586 10.39 0.0016

Error 130 2.35336747 0.01810283

Corrected Total 131 2.54153333

R-Square Coeff Var Root MSE PSS_2013 Mean

0.074036 63.06879 0.134547 0.213333

Source DF Type I SS Mean Square F Value Pr > F

H3 1 0.18816586 0.18816586 10.39 0.0016

Source DF Type III SS Mean Square F Value Pr > F

H3 1 0.18816586 0.18816586 10.39 0.0016

Contrast DF Contrast SS Mean Square F Value Pr > F

H3 1 0.18816586 0.18816586 10.39 0.0016

Standard

Parameter Estimate Error t Value Pr > |t|

Intercept 0.2281706052 0.01258261 18.13 <.0001

H3 -.1357147781 0.04209494 -3.22 0.0016

----------------------------------------------------------------------------------------------------

Test for Haplotype 4 16:33 Monday, February 29, 2016 29

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Number of Observations Read 132

Number of Observations Used 132

----------------------------------------------------------------------------------------------------

Test for Haplotype 4 16:33 Monday, February 29, 2016 30

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Dependent Variable: PSS_2013

Sum of

Source DF Squares Mean Square F Value Pr > F

Model 1 0.12443421 0.12443421 6.69 0.0108

Error 130 2.41709913 0.01859307

Corrected Total 131 2.54153333

R-Square Coeff Var Root MSE PSS_2013 Mean

0.048960 63.91707 0.136356 0.213333

Source DF Type I SS Mean Square F Value Pr > F

H4 1 0.12443421 0.12443421 6.69 0.0108

Source DF Type III SS Mean Square F Value Pr > F

H4 1 0.12443421 0.12443421 6.69 0.0108

Contrast DF Contrast SS Mean Square F Value Pr > F

H4 1 0.12443421 0.12443421 6.69 0.0108

Standard

Parameter Estimate Error t Value Pr > |t|

Intercept 0.2242132838 0.01259143 17.81 <.0001

H4 -.1263698272 0.04884826 -2.59 0.0108

----------------------------------------------------------------------------------------------------

Test for Haplotype 5 16:33 Monday, February 29, 2016 31

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Number of Observations Read 132

Number of Observations Used 132

----------------------------------------------------------------------------------------------------

Test for Haplotype 5 16:33 Monday, February 29, 2016 32

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Dependent Variable: PSS_2013

Sum of

Source DF Squares Mean Square F Value Pr > F

Model 1 0.06560886 0.06560886 3.44 0.0657

Error 130 2.47592448 0.01904557

Corrected Total 131 2.54153333

R-Square Coeff Var Root MSE PSS_2013 Mean

0.025815 64.69017 0.138006 0.213333

Source DF Type I SS Mean Square F Value Pr > F

H5 1 0.06560886 0.06560886 3.44 0.0657

Source DF Type III SS Mean Square F Value Pr > F

H5 1 0.06560886 0.06560886 3.44 0.0657

Contrast DF Contrast SS Mean Square F Value Pr > F

H5 1 0.06560886 0.06560886 3.44 0.0657

Standard

Parameter Estimate Error t Value Pr > |t|

Intercept 0.2208798843 0.01268136 17.42 <.0001

H5 -.0873881772 0.04708347 -1.86 0.0657

----------------------------------------------------------------------------------------------------

Test for Haplotype 6 16:33 Monday, February 29, 2016 33

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Number of Observations Read 132

Number of Observations Used 132

----------------------------------------------------------------------------------------------------

Test for Haplotype 6 16:33 Monday, February 29, 2016 34

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Dependent Variable: PSS_2013

Sum of

Source DF Squares Mean Square F Value Pr > F

Model 1 0.00496349 0.00496349 0.25 0.6149

Error 130 2.53656985 0.01951208

Corrected Total 131 2.54153333

R-Square Coeff Var Root MSE PSS_2013 Mean

0.001953 65.47764 0.139686 0.213333

Source DF Type I SS Mean Square F Value Pr > F

H6 1 0.00496349 0.00496349 0.25 0.6149

Source DF Type III SS Mean Square F Value Pr > F

H6 1 0.00496349 0.00496349 0.25 0.6149

Contrast DF Contrast SS Mean Square F Value Pr > F

H6 1 0.00496349 0.00496349 0.25 0.6149

Standard

Parameter Estimate Error t Value Pr > |t|

Intercept 0.2154476441 0.01286049 16.75 <.0001

H6 -.0252461592 0.05005571 -0.50 0.6149

----------------------------------------------------------------------------------------------------

Test for Haplotype 7 16:33 Monday, February 29, 2016 35

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Number of Observations Read 132

Number of Observations Used 132

----------------------------------------------------------------------------------------------------

Test for Haplotype 7 16:33 Monday, February 29, 2016 36

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Dependent Variable: PSS_2013

Sum of

Source DF Squares Mean Square F Value Pr > F

Model 1 0.04871084 0.04871084 2.54 0.1134

Error 130 2.49282249 0.01917556

Corrected Total 131 2.54153333

R-Square Coeff Var Root MSE PSS_2013 Mean

0.019166 64.91055 0.138476 0.213333

Source DF Type I SS Mean Square F Value Pr > F

H7 1 0.04871084 0.04871084 2.54 0.1134

Source DF Type III SS Mean Square F Value Pr > F

H7 1 0.04871084 0.04871084 2.54 0.1134

Contrast DF Contrast SS Mean Square F Value Pr > F

H7 1 0.04871084 0.04871084 2.54 0.1134

Standard

Parameter Estimate Error t Value Pr > |t|

Intercept 0.2071830763 0.01265543 16.37 <.0001

H7 0.0721643793 0.04527767 1.59 0.1134

----------------------------------------------------------------------------------------------------

Test for Haplotype 8 16:33 Monday, February 29, 2016 37

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Number of Observations Read 132

Number of Observations Used 132

----------------------------------------------------------------------------------------------------

Test for Haplotype 8 16:33 Monday, February 29, 2016 38

------------------------------------------------------------- Window=1 -------------------------------------------------------------

The GLM Procedure

Dependent Variable: PSS_2013

Sum of

Source DF Squares Mean Square F Value Pr > F

Model 1 0.52691021 0.52691021 34.00 <.0001

Error 130 2.01462312 0.01549710

Corrected Total 131 2.54153333

R-Square Coeff Var Root MSE PSS_2013 Mean

0.207320 58.35345 0.124487 0.213333

Source DF Type I SS Mean Square F Value Pr > F

H8 1 0.52691021 0.52691021 34.00 <.0001

Source DF Type III SS Mean Square F Value Pr > F

H8 1 0.52691021 0.52691021 34.00 <.0001

Contrast DF Contrast SS Mean Square F Value Pr > F

H8 1 0.52691021 0.52691021 34.00 <.0001

Standard

Parameter Estimate Error t Value Pr > |t|

Intercept 0.1932524844 0.01136935 17.00 <.0001

H8 0.2484527632 0.04260894 5.83 <.0001

====================================================================================================

Re: Haplotype Regression

Hi,

I must correct myself when I said LS Means in the previous reply; instead, the values I defined as LS Means are indeed the predicted means for each haplotype.

My apologies.

Luciano Silva

Re: Haplotype Regression

Hi,

If I understood your question correctly, you want the Least Square Means (LS Means) of each haplotype. Since you have results from the Overall Test and Individual Test, you can get two LS Means estimates for each haplotype. In either case, the LS Means for the haplotype is the sum of the Intercept and the haplotype estimate. The LS Means from the Overall Test is conditional on other haplotypes fitted together in the model and it may change if you change the model (say, remove haplotypes from the model), while the LS Means from the Individual Test is unconditional to other haplotypes because the model has only one haplotype.

LS Means H1 (overall test) = 0.2890451760 + (-.0302601288) = 0.2587850472

LS Means H1 (individual test) = 0.2058495133 + 0.0604036959 = 0.2662532092

Sincerely yours,

Luciano Silva

Re: Haplotype Regression

Hi,

Thanks for posting this question.

First, if you are looking for the mean for your trait (PSS_2013) for each haplotype, then take a look at each section labelled Test for Haplotype Hx where x is the Haplotype number. Within each section there is an item called PSS_2013 Mean. That would be the mean for the trait for that haplotype. Is that what you are looking for? These individual sections for each haplotype are the individual tests for each haplotype as opposed to the whole model fit where all haplotypes are fit together (at the beginning) and the PSS_2013 Mean is the mean for the whole fit, instead of the individual fits.

The Parameter Estimates are the slope of the fit for each haplotype as individual tests or the whole model fit and represent the effect (+ is a positive effect and - is a negative effect) on that trait for each haplotype or when fit together, their effect when considering the other haplotypes.

Does that help explain the output?

Best,

Chris Kirchberg

Comments

There are no comments yet