The Sibship Index (SI) can help identify the sibling relationship between 2 persons, separating full from half siblings. This method is based on two person phenotypes and Bayes's theorem.
sibship index =
= (probability two persons truly related) / (probability two persons unrelated)
Genotype Person #1 |
Genotype Person #2 |
Formula for Sibship Index |
AA |
AA |
(k2 + (2 * k1 * a) + (k0 * (a^2))) / (a^2) |
AA |
AB |
(k1 + (k0 * a)) / a |
AA |
BB |
k0 |
AA |
BC |
k0 |
AB |
AB |
(k2 + (k1 * a) + (k1 * b)+ (k0 * 2 * a * b)) / (2 * a * b) |
AB |
AA |
(k1 + (k0 * a)) / a |
AB |
AC |
(k1 + (2 * k0 * a)) / (2 * a) |
AB |
CD |
k0 |
Sibship |
full |
half |
k0 |
0.25 |
0.50 |
k1 |
0.25 |
0.25 |
k2 |
0.25 |
0 |
where:
• a = frequency of allele A
• b = frequency of allele B
ratio of full to half sibship =
= (SI for full sibling) / (SI for half sibling)
Interpretation:
• If the full SI > half SI (ratio > 1), then the 2 people are probably full siblings. In the series of Wenk (2000), 72% of full siblings and 8% of half siblings had a ratio > 1.0.
• If the full SI < half SI (ratio < 1), then the 2 people are probably half siblings. In the series of Wenk (2000), 28% of full siblings and 92% of half siblings had a ratio < 1.0.
• Examination of several alleles may be necessary to determine the sibship of 2 persons. Full siblings will share 2 alleles at several loci (about 25% of the total), while half siblings will share 2 alleles only rarely or not at all.
Specialty: Genetics
ICD-10: ,