Comparing genome results from 23andMe and deCODEme

Antonio C B Oliveira

acbolive at sloan.mit.edu

April 2008

In order to determine the accuracy of the genome profile provided by 23andMe and deCODEme I arranged to be genotyped by both companies and wrote a computer program to compare the results.

The downloaded files contains 576105 snips in the case of 23andMe and 1013349 snips for deCODE. After removing the no-call and matching the two files by SNP identification, 560299 snips were present in both files. The comparisson revealed 23 cases in which the results do not agree. The discrepancy by chromosome is summarized in the following table.


Chrm	equal	not equal
1	42060	2	rs4660646	rs1934496
2	45328	1	rs10933181
3	37637	1	rs9881405
4	33557	3	rs1064205	rs312330	rs11100437
5	34530	0
6	36742	1	rs4458717
7	30199	0
8	31736	1	rs2955195
9	26745	3	rs7033246	rs1536928	rs10793963
10	29187	0
11	27367	3	rs10894749	rs3921012	rs510978
12	27092	1	rs12296276
13	20766	0
14	18356	0
15	16630	2	rs754499	rs4965862
16	16863	1	rs11149566
17	14391	1	rs2290505
18	16799	1	rs12960185
19	9512	0
20	14240	1	rs4814138
21	8193	0
22	8443	0
MT	0	0
X	13793	1	rs6615048
Y	110	0
Total	560276	23

This error rate seems to me to be quite acceptable and I wonder if this is the rate expected in scientific studies using the same technology.