Dear Methee,

Based on your question, most probably your study contains two groups (case and control) and the aim of study is to examine if the allele or genotype frequencies of each SNP is significantly different between your case and control groups or not. If we suppose that the design of your study is exactly as it mentioned above, a chi-square test  can answer your hypothesis. 

In your study, if you aim to investigate the relationship between your groups (case and control) and allele frequency (Allele 1 and 2), you can apply this test. You can also use this test to assess the relationship between your groups and different genotypes (11, 12 and 22).

The point is that in chi-square test, one (or both) variables may have more than two levels, and that the variables do not have to have the same number of levels. In your study, the group has two levels (case and control); however, the genotypes have three levels.

To see how you can run your data in SPSS, please check the attached file.

In addition, please remember that the chi-square test assumes the expected value of each cell is five or higher. However, if this assumption is not met in your data, you can use Fisher's exact test.

1-Go to analyze >>> Descriptive statistics

2- Go to Frequencies. Move the SNP into the box, and then press OK

If you want to detect association between SNPs using the chi-square test:

1- Go to analyze>>> Descriptive  Statistics

2- Go To Cross tabulation, put one variable in rows and the other in the columns,  click on STATISTICS, and choose Chi-square

Dear Methee,

Although I'm not aware of the exact nature of your work, here are some general suggested informations which can be used if you are abour to determine ORs for a certain condition comparing patients wit age- and sex-matched controls (and please bear in mind that controls should be age-matched within at a maximum of 18 months interval of the date of birth. The Hardy-Weinberg frequencies for all alleles in patients and controls should be analyzed using exact probability tests available in Mendel software (V5.7.2).The Kolmogorov-Smirnov and Shapiro-Wilk tests should be used to verify the normality of continuous variables and the Levene test to analyze the homogeneity of variances. Differences in genotype frequency, age class and gender distributions between patients and controls should be evaluated by the χ2 test. Adjusted odds ratio (OR) and corresponding 95% confidence interval (CI) should be calculated using unconditional multiple logistic regression. The model for adjusted OR should  include terms for sex, age at diagnosis (e.g., ≤30, 31-39, 40-49, 50-59, 60-69 and ≥70 years). The reference groups for these variables should be clearly set. All analyses can be performed using SPSS 15.0 (SPSS, Inc.).

Jose Rueff

Dear Methee,

The attached file is a manual about how to calculate odds ratio using SPSS. Hope you find it helpful.

Thank you very much  :) for your valuable comments and suggestion.

Thanks for sharing your views; and I also agree with your advice.

thanks very much for your advice