Clipping  markers which are located far from the iterated locus

 

As the result of the experiments that were performed,  we found that clipping  until half of the total number of markers doesn't influence the LOD score when we clip each time the marker farthest from the iterated locus. We also found that when the iterated locus is the first or the last  on the map, we can clip farthest markers until 3/4 of the total number of markers without influence on the LOD score. The maximum distance from the varied locus to markers that influence the final LOD score is various for different input data files, as well as the number of those markers.

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

 

datafile40_70_9.dat, pedfile40_70_9.dat, Varied Locus = 20

Num Loci

Num People

Exact LOD score

Approx LOD score

error rate

ratio

40

70

-14.013844

-14.013844

0%

19/40

40

70

-14.013844

-17.382109

24%

23/40

40

70

-14.013844

-18.662389

33%

26/40

 

datafile40_70_9_1.dat, pedfile40_70_9_1.dat, Varied Locus = 1

Num Loci

Num People

Exact LOD score

Approx LOD score

error rate

ratio

40

70

-2.331068

-2.331068

0%

29/40

40

70

-2.331068

-2.259107

3%

34/40

40

70

-2.331068

-2.781365

19.3%

36/40

 

datafile76_20_7.dat, pedfile76_20_7.dat, Varied Locus = 38

Num Loci

Num People

Exact LOD score

Approx LOD score

error rate

ratio

76

20

-11.597888

-11.597888

0%

37/76

76

20

-11.597888

-13.214128

13.9%

45/76

76

20

-11.597888

-13.349134

15%

58/76

 

datafile76_20_7_1.dat, pedfile76_20_7_1.dat, Varied Locus = 1

Num Loci

Num People

Exact LOD score

Approx LOD score

error rate

ratio

76

20

-2.885523

-2.885523

0%

56/76

76

20

-2.885523

-2.885523

0%

65/76

76

20

-2.885523

-2.819683

2%

73/76

 

The following graphs represent the  LOD error rate as a function of  the percent of clipped markers (farthest markers). Error rate increases when the number of clipped markers grows up. The way it increases is various for different input files. However, it can be seen that clipping lower than 50% of total number of markers, in the case of  "iterated locus is in the middle of the map", and lower that 75% of total number of markers , in the case of "iterated locus is the first or the last on the map",  gives LOD error rate = 0 for different input files. Each graph represents LOD error rate for two different input files according to experimental results mentioned above.

To see the graph for the case of "iterated locus is in the middle of the map" click here.

To see the graph for the case of "iterated locus is the first or the last on the map" click here.

 

 

Back to main page