Clipping less informative markers

 

As the result of the experiments that were performed,  we  found that the  percent of non-information of the marker ( percent of people that are unknown in this locus)  influences  the LOD score as well as the proximity of the marker to the iterated locus. The lower  the percent of  non-information of the marker  and  the proximity  to the iterated locus are, the bigger is the influence of this marker on the LOD score. (By proximity we mean a ratio that is defined as: ratio=(the distance from the marker to the varied locus)/(The distance from varied locus to the farthest locus on the map).)   We defined  bounds for the percent of non-information (PERCENT_BOUND) and for the ratio (INFO_DISTANCE_BOUND), above which clipping markers doesn't influence the LOD score.

Experimental results for measuring the influence of clipping less informative markers on the LOD score:

 

datafile20_15_5.dat, pedfile20_15_5.dat, PERCENT_BOUND = 0.4

Num Loci

Num People

Exact LOD score

Approx LOD score

error rate

INFO_DISTANCE_BOUND

20

20

-0.161837

-0.162201

0.2%

0.55

20

20

-0.161837

-0.187914

16%

0.35

20

20

-0.161837

-0.113102

30.1%

0.2

 

datafile20_20_7.dat, pedfile20_20_7.dat, PERCENT_BOUND = 0.4

Num Loci

Num People

Exact LOD score

Approx LOD score

error rate

INFO_DISTANCE_BOUND

20

20

0.042357

0.042879

1.2%

0.4

20

20

0.042357

0.043108

1.7%

0.3

20

20

0.042357

-0.494675

1068%

0.2

 

 

 

datafile20_20_7.dat, pedfile20_20_7.dat, INFO_DISTANCE_BOUND = 0.4

Num Loci

Num People

Exact LOD score

Approx LOD score

error rate

PERCENT_BOUND

20

20

0.042357

0.042345

0.02%

0.65

20

20

0.042357

0.042268

0.2%

0.6

20

20

0.042357

-0.043108

1.8%

0.5

 

datafile20_15_5.dat, pedfile20_15_5.dat, INFO_DISTANCE_BOUND = 0.4

Num Loci

Num People

Exact LOD score

Approx LOD score

error rate

PERCENT_BOUND

20

15

-0.161837

-0.161837

0%

0.6

20

15

-0.161837

-0.161831

0%

0.5

20

15

-0.161837

-0.187914

16%

0.4

 

Back to main page