1
$\begingroup$

I have a SNP stats file structure, which contains all information about genotypes and imputed SNP/INDEL imputation qualities, allele frequencies and minor allele assignment.

SNPID   RSID    chromosome  position    A_allele    B_allele    minor_allele    major_allele    AA  AB  BB  AA_calls    AB_calls    BB_calls    MAF HWE missing missing_calls   information
--- rs200405949 NA  10023   CCAA    C   C   CCAA    923.32  16.662  0.012   911 6   0   0.0088756   4.8216e-17  1.0638e-05  0.024468    0.47987
--- rs201803828 NA  10097   CCA C   C   CCA 930.29  9.711   0   924 4   0   0.0051654   -0  1.0638e-06  0.012766    0.50802
--- rs185444096 NA  10177   C   T   T   C   291.03  45.234  0.712   216 12  0   0.069231    9.6433e-17  0.64152 0.75638 0.80608
--- rs200882690 NA  10299   TA  T   T   TA  292.62  43.613  0.744   219 12  0   0.066921    4.8216e-17  0.64152 0.75426 0.79622
--- rs56377469  NA  10469   G   C   C   G   239.25  474.83  225.9   38  69  20  0.4929  0.10014 2.234e-05   0.86489 0.42593
--- rs7341907   NA  10869   C   G   G   C   239.29  474.8   225.89  38  69  20  0.49287 0.10014 2.234e-05   0.86489 0.42597
--- rs149305563 NA  14665   G   A   A   G   267.44  66.53   2.973   144 16  0   0.10755 0.10741 0.64155 0.82872 0.79468
--- rs149079262 NA  14690   C   G   G   C   207.26  115.51  14.189  56  25  1   0.2135  0.12672 0.64153 0.9117  0.79388
--- rs141156662 NA  15883   A   G   G   A   259.05  71.908  6.001   131 13  1   0.12451 0.20633 0.64154 0.84468 0.80157
--- rs202192731 NA  17614   CT  C   CT  C   7.893   142.55  789.45  1   39  581 0.084231    0.27659 0.0001117   0.33511 0.47425

I managed to assume/figure out the meaning of all columns but two - HWE and information.

HWE: I assume HWE means Hardy–Weinberg equilibrium, and is actually a chi-square value obtained via Pearson's chi-squared test, and using only samples from the same dataset (and not compared to some general dataset, e.g. GWAS). Using HWE , we can then calculate the p-value using the degrees of freedom (one in this case) via the gamma function (e.g. via p-value calculator), and if the result is significant, this means that the specific SNP is not in HWE, hence this SNP is either in linkage disequilibrium, or needs to be rejected/ignored because of genotyping (i.e. data obtaining) errors. If so, what significance level is mostly used/best for such cases? In the Hapmap II paper, SNPs are removed if they are not in HWE with a p-value < 0.001.

information: I have no clue what it means or how it's calculated.

$\endgroup$
  • $\begingroup$ Regarding "information": specifying the software used for analysis could be of help. As of now, I have no idea of what it is. Regarding "HWE": 1) Not knowing the software I am not sure, but I can guess that In the column you already have the pvalue. 2) There is no clear agreement on the threshold, and not even if SNPs outside of HWE should be removed or not. You can give a look to this paper (ncbi.nlm.nih.gov/pmc/articles/PMC2796342) to start understanding. The choice depends a lot on the study sample and the questions you are trying to answer. $\endgroup$ – Fabio Marroni Feb 19 at 16:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.