Skip to main content

Table 3 Comparisons between the proposed method and Park's method.

From: Shrunken methodology to genome-wide SNPs selection and construction of SNPs networks

   The Proposed Method Park's Method
Chromosome Number of SNPs Accuracy Δ Number of SNPs Accuracy Δ Number of SNPs
1 29226 0.5926 0.893 21 0.5667 0.876 88
2 30298 0.6019 0.919 20 0.5722 0.965 87
3 25648 0.6056 0.880 26 0.5611 0.860 132
4 22315 0.6204 0.918 11 0.5945 1.024 42
5 22746 0.5796 0.926 22 0.5611 0.900 82
6 24334 0.6093 1.003 3 0.5204 0.944 49
7 19740 0.5797 0.901 7 0.5685 0.980 39
8 21384 0.5834 0.931 6 0.5741 0.979 35
9 18122 0.5426 0.932 7 0.5148 0.908 61
10 18525 0.6074 0.930 9 0.5593 0.999 47
11 18074 0.6352 0.964 7 0.5944 0.942 52
12 18186 0.6074 0.893 17 0.5667 0.975 31
13 13077 0.5870 0.906 9 0.5053 0.919 23
14 11728 0.5574 0.901 8 0.5259 0.962 26
15 10813 0.6037 0.905 8 0.5556 0.917 40
16 10892 0.5778 0.936 8 0.5611 0.946 37
17 10730 0.5815 0.919 4 0.5037 0.961 31
18 11677 0.5704 0.921 5 0.5071 0.956 27
19 7749 0.6037 0.887 7 0.5463 0.950 13
20 9647 0.6111 0.864 12 0.5907 0.951 33
21 6070 0.5833 0.875 8 0.4945 0.911 8
22 6459 0.6056 0.901 5 0.5444 0.924 23
Average 0.5930 0.914 10 0.5495 0.943 46
  1. Illustration of the classification accuracies of all chromosomes of Parkinson disease genome-wide data set after a 10-fold cross validation and a detailed comparison between our proposed method and Park's method when filter out those SNPs whose missing percentage is >1%. The second column indicates the number of SNPs in each chromosome after filtering. Accuracy is the average accuracy after a 10-fold cross validation, Δ value and No of SNPs are obtained when the corresponding average accuracy is reached.