Table 4.
Computational time required by 2LOmb to analyse large-scaled simulated data sets with different numbers of available SNPs, different sample sizes and different ratios of case samples from two affected groups where the affected status is governed by a two-locus interaction
|
Sample |
Ratio of |
Computational time (sec) |
|
|---|---|---|---|
| size | case samples | 10,000 SNPs | 100,000 SNPs |
| 1,600 |
1:3 |
34 |
3,106 |
| |
1:1 |
34 |
3,116 |
| 3,200 |
1:3 |
68 |
6,227 |
| |
1:1 |
68 |
6,256 |
| 6,400 |
1:3 |
135 |
12,503 |
| 1:1 | 136 | 12,560 | |
The simulation is carried out on a computer system with a graphics processing unit. The parallelism of the graphics processing unit is exploited to speed up the computation. The computer system is equipped with an AMD 2.8 GHz quad-core processor, 4GB of main memory and an NVIDIA GeForce GTX 285 graphics processing unit. The graphics processing unit contains 240 streaming processors sharing 1GB of GDDR3 memory. Each streaming processor has a clock rate of 1.48 GHz. An Ubuntu 9.10 operating system is installed on the computer system. The computational time is collected from the processing of multiple independent data sets for each simulation setting. The displayed time is the maximum time required to analyse one data set.