How to obtain likelihood ratios using SNPs

Summary of this article

Learn how to calculate likelihood ratios using SNPs. Learn the basics of increasing the probability of parent-child or blood relationship based on genetic data.

To find the likelihood ratio using SNPs, the probability of genotype match is calculated, which is then used to calculate the likelihood for a specific hypothesis (e.g. the hypothesis that a person has a specific parent-child relationship).Specifically, the following steps are performed.

SNP data collection: collect SNP data from the DNA of the parent-offspring candidates.
Match probability calculation: calculates the SNP match probability between parent-offspring candidates.
Likelihood ratio calculation: computes the likelihood ratio by comparing the likelihoods of the parent-offspring and non-parent-offspring relationship hypotheses.

For example, if the probability that the SNPs match is p and the probability that they do not match is q, the likelihood ratio (LR) is expressed as p/q.

Specific examples are given here.

Assumptions:
Allele of SNP A: parent ‘AA’, child ‘AA’.
Frequency of allele A in population:

1. likelihood in the presence of a parent-child relationship:

If the parent is ‘AA’, the probability that the child is also ‘AA’ is 1.

2. likelihood of no parent-child relationship:

The probability that a child is ‘AA’ depends on how common allele A is in the population.

3. calculation of the likelihood ratio (LR):

The likelihood ratio of 100 indicates that the hypothesis of a parent-child relationship is 100 times more favored than the hypothesis of no parent-child relationship.

If multiple SNPs are used, a similar calculation is performed for each SNP and their likelihood ratios are multiplied to obtain the final likelihood ratio.