Home
Download
Getting started
Manual
Contact
Help
This document provides a
step-by-step description of how to use FamLinkX to
calculate LR in two different scenarios. The first case considers a case of two
maternal half siblings versus unrelated using data from two linked STR markers.
The second case includes data from a cluster of three SNP markers on a case of
two male maternal cousins versus unrelated.
First consider a cluster of two
STR X-markers, L1 located on 10 cM and L2 located on 10.05 cM. (You can skip the first step, defining the
database, by opening example1.sav supplied in the
install directory of FamLinkX)
1.
Open FamLinkX
2.
Click on the “Frequency database” option in the
File menu.
3.
Click the “Add” button to add a new cluster
4.
Double click the created cluster or
alternatively select and click the “Edit” button
5.
Add two new allele systems by clicking the “Add”
button in the Allele system section twice.
6.
Define the system, name L1, set
genetic position to 10.
7.
Now add four alleles, 13, 14, 15 and 16 with
frequencies 0.1, 0.1, 0.1 and 0.7.
8.
Define mutation parameters.
a.
Click the “Mutations” button
b.
Make sure the Extended step wise model is
selected for both male and female mutations. The rate should be 0.005, the
range 0.1 and the second rate (rate2) 0.000001
9.
Repeat steps 6->8 for the second system, L2,
use the same alleles and mutation parameters but change the genetic position to
10.05
10.
Now add observed haplotypes by clicking the
“Add” button in the Observed haplotypes section.
a.
Add one haplotype with 10 observations, name it
haplo1. The haplotype configuration should be [13 13]
11.
Last, close all windows and save the project,
File menu -> Save.
We have now defined the
database and will continue by computing likelihoods for a case. We will
consider a case of two female maternal half-siblings.
1.
Click File -> New wizard or Tools ->
Select pedigree
2.
Select Half Siblings (Maternal
) and click Next.
3.
Next select alternative hypothesis. Select
Unrelated and click next.
4.
Enter DNA data by select person 1 (1. NN) from
the drop down list.
5.
Change name to Person1 and make sure gender is
set to female.
6.
Select Cluster1 from the drop down list.
7.
Then for each system enter DNA data by changing
all alleles to 13.
8.
Repeat steps 4-7 for person2.
9.
The two persons should now be homozygous for
alleles 13 at both defined allele systems.
10. Click
next and click on the “Calculate” tab. The likelihood for the hypotheses,
and the corresponding LR, is calculated. When you select the ‘Half Siblings (Maternal)’ pedigree you should see the output :
LR(Exact):
1.02858
LR(Cluster): 1.04945
LR(LE):
50.4595
11. The likelihoods are calculated using three different models; LR(Exact) which is the model presented by Kling et al., LR(Cluster) which is the model provided for clusters of markers in Merlin and LR(LE) which is the model for linked markers in Merlin. The last two models cannot handle mutations which in the current case setup is irrelevant. Try changing the alleles such that a mutation is necessary to explain the data. Only LR(Exact) should then have a non-zero LR.
12.
The result can be stored for instance by
clicking ‘Save Results’, select
case report and Save.
The second example involves
three SNP markers defined in one cluster. (You can skip the first step,
defining the database, by opening example2.sav
supplied in the install directory of FamLinkX)
1.
Open FamLinkX.
2.
Click on the “Frequency database” option in the
File menu.
3.
Click the “Add” button to add a new cluster
4.
Double click the created cluster or
alternatively select and click the “Edit” button
5.
Add three new SNP:s by
clicking the “Add” button in the Allele system section thrice.
6.
Define the system, name S1, set
genetic position to 10.
7.
Now add two alleles, 1 and 2 with frequencies
0.4 and 0.6. (It is currently not possible to define alphabetic alleles and
SNP:s therefore has to be defined as numerical alleles)
8.
Define mutation parameters.
a.
Click the “Mutations” button
b.
Make sure the Extended step wise model is
selected for both male and female mutations. The rate should be 0.000001, the
range 0.1 and the second rate (rate2) 0.000001 (For SNP:s we are not really interested
in mutations in the same sense as for STR markers as mutations for SNP:s do not
follow a step-wise mechanism)
9.
Repeat steps 6->8 for the second and third
SNP:s, S2 and S3, use the same alleles and mutation parameters but change the
genetic positions to 10.1 and 10.2
10.
Now add observed haplotypes by clicking the
“Add” button in the Observed haplotypes section.
a.
Add one haplotype with 30 observations, name it
haplo1. The haplotype configuration should be [1 1 1]
b.
Add haplotypes [1 1 2] to [2 2 2] using the following number of observations (name them
haplo2, haplo3 etc.)
|
S1 |
S2 |
S3 |
Observations |
|
1 |
1 |
2 |
20 |
|
1 |
2 |
1 |
5 |
|
1 |
2 |
2 |
5 |
|
2 |
1 |
1 |
5 |
|
2 |
1 |
2 |
5 |
|
2 |
2 |
1 |
20 |
|
2 |
2 |
2 |
10 |
11.
Last, close all windows and save the project,
File menu -> Save.
We have now defined the
database and will continue by computing likelihoods for a case. We will
consider a case of two male maternal cousins.
1.
Click File -> New wizard or Tools ->
Select pedigree
2.
Scroll down and select Maternal Cousins and
click Next.
3.
Next select alternative hypothesis. Select
Unrelated and click next.
4.
Enter DNA data by select person 1 (1. NN) from
the drop down list.
5.
Change name to Person1 and make sure gender is
set to male.
6.
Select Cluster1 from the drop down list.
7.
Then for each system enter DNA data by changing
alleles to 1, 2 and 2 for systems S1, S2 and S3 respectively.
8.
Repeat steps 4-7 for person2.
9.
The two males should both have the haplotype [1
2 2].
10. Click
next and click on the “Calculate” tab. The likelihood for the
hypotheses, and the corresponding LR, is calculated. When you select the ‘Maternal cousins’ pedigree you should see the output :
LR(Exact):
7.9718
LR(Cluster): 7.98795
LR(LE):
3.22441
11. The likelihoods are calculated using three different models; LR(Exact) which is the model presented by Kling et al., LR(Cluster) which is the model provided for clusters of markers in Merlin and LR(LE) which is the model for linked markers in Merlin. The last two models cannot handle mutations which in the current case setup is irrelevant since mutations cannot be observed. (Though worth noting is that two alleles can be identical by descent without being identical by state using our model)
The result can be stored for instance by clicking ‘Save Results’, select case report and
Save. A
more comprehensive description of all the features of FamLinkX,
and more examples, can be found in the manual.
You may send comments to help@famlink.se