EASYPAT 29.08.2000
EASYPAT is a computer program that allows calculation of likelihood
ratios for single locus data, comparing specific types of simple
hypotheses regarding the familial relationships involved.
Currently, these hypotheses include
1. Parenthood vs. non-Parenthood in duos comprising parent and
child (DIRECT=2). The program also calculates the likelihood under
the hypothesis that, instead of the putative parent, a first degree
relative is the biological parent.
2. Paternity vs. non-paternity in trios comprising mother,
child and alleged father (DIRECT=3). The likelihood is also given
under the hypothesis that a first degree relative of the alleged
father is the biological father.
3. Full sibs vs. halfsibs in trios comprising one parent and
two children (DIRECT=4). In this case, the likelihood is also
calculated assuming that the unknown parents are first degree
relatives.
EASYPAT requests an input file of the following format:
1. line; DIRECT, a single positive integer that identifies the
types of hypotheses compared (see above).
2. line; NMARK, a single positive integer indicating the total
number of markers tested. All individuals must be typed for all
markers included.
Following lines; for each marker, enter the name of the marker
and the genotypes of the respective individuals in the order:
DIRECT=2 - child, putative parent
DIRECT=3 - mother, child, putative father
DIRECT=4 - parent, 1. child, 2. child
All entries must be separated by at least one blank. Therefore,
blanks must not be part of marker or allele designations!
3
5
THO1 199 199 196 199 192 196
D19 210 234 230 234 230 234
CSF 311 315 311 311 311 315
VWA 152 156 152 160 140 160
CYAR 177 177 177 196 196 196
In this example, mother, child and putative father have been typed
at five markers.
The allele frequencies and mutation rates necessary to calculate
likelihoods are enquired by the program during execution. Entering
the following allele frequencies for the above example
THO1 196 : 0.30
D19 230 : 0.20
CSF 311 : 0.15
VWA 160 : 0.15
CYAR 196 : 0.25
yields the following output
Paternity vs. Non-paternity
Likelihood ratio : 1.9841269841E+02
W value : 99.49853%
First degree relative
Likelihood ratio : 2.8888888889E+01
W value : 96.65428%
apriori posteriori [1.degree relative]
-----------------------------------------
0.1 95.66082% 76.24633%
0.2 98.02384% 87.83784%
0.3 98.83767% 92.52669%
0.4 99.24967% 95.06399%
0.5 99.49853% 96.65428%
0.6 99.66513% 97.74436%
0.7 99.78447% 98.53817%
0.8 99.87416% 99.14204%
0.9 99.94403% 99.61686%
-----------------------------------------
Alternatively, allele frequencies can be loaded from an
input file of the following format
1. line; SNMARK, a single positive integer equal to the number
of markers for which allele frequencies are to be stored.
The following will be repeat SNMARK times
2. line; marker designation and number of alleles (NALL), separated by
at least one blank. The marker designation must start at
the first position of this line. Do not include leading blanks.
NALL lines; allele designation and allele frequency, separated by
at least one blank. The allele designation must start at
the first position of this line. Do not include leading blanks.
3
D2S367 4
141 0.40134
139 0.33779
115 0.23311
143 0.02776
D12S66 3
167 0.37841
171 0.26809
163 0.35350
D8S601 5
228 0.37703
224 0.21486
226 0.16689
230 0.13446
222 0.10676
Please note that marker and allele designations in the two types
of input file must be IDENTICAL, otherwise the program will not be
able to identify alleles properly. Frequencies of unidentifiable
alleles will be enquired during program execution.
Computational methodology employed in EASYPAT
- Trios (DIRECT=3)
The likelihood of paternity, L1, is calculated by considering
the four possible inheritance patterns of the parents. L1 is
increased by 1/4 for each pattern that is compatible with
the offspring genotype.
The likelihood of non-paternity, L2, is calculated considering
only the two possible inheritance patterns of the mother.
If a pattern is compatible with the offspring genotype,
this defines a paternal allele, say A, and L2 is increased
by half the population frequency, f(A)/2, of this allele.
If the paternal genotype is not compatible with the offspring
genotype, i.e. if the putative father does not carry any of
the necessarily paternal alleles, then L1 is set equal to the
overall mutation rate and L2 is set equal to either
1-f(A), if the genotype of the putative father is AA, or
1-f(A)-f(B), if the genotype of the putative father is AB.
The information used in this instance is thus confined
to the fact that the paternal allele of the child is different
from the allele(s) carried by the putative father. Implicitly,
it is also assumed that each allele is equally likely to mutate.
If the maternal genotype is not compatible with the offspring
genotype, the program is aborted.
The likelihood, L3, of a first degree relative of the putative
father being the biological father is calculated as (L1+L2)/2.
(Half of the genetic material of the two is identical, the
other half is unrelated).
- Duos (DIRECT=2)
The likelihood of parenthood, L1, is the same as L2 is for
trios. The likelihood of non-paternity, L2, equals the population
frequency of the offspring genotype under Hardy-Weinberg
conditions. If the parental genotype is not compatible with the
offspring genotype, L1 is set equal to the overall mutation rate
while L2 is set equal to 1-f(A), if the putative parent is AA,
or 1-f(A)-f(B), if the putative parent is AB (see above). The
likelihood L3 of parenthood by a 1. degree relative is again
calculated as (L1+L2)/2.
- Trios (DIRECT=4)
The likelihoods of full sibship, L1, and half-sibship, L2, are
calculated considering the four possible segregation patterns
of the typed parent. If a pattern is compatible with the two
offspring genotypes, this defines alleles A and B that must have
been inherited from the untyped parent(s). If A is not equal to
B then L2 is increased by f(A)*f(B)/4, and L1 is increased by
f(A)*f(B)/8. If A=B, then L2 is increased by f(A)*f(A)/4 and
L1 is increased by f(A)*[1+f(A)]/8. The likelihood L3 in this
case refers to the hypothesis that the untyped parents of
the two offspring were 1. degree relatives.
Best wishes
MK