Regardless of allele matching at markers, the PI value depends on the distinctiveness of the matching. If it is known that there are many people that match alleles at the same location, then the PI value would be much less, like 0.0782. So just because the two subjects match at the same location with the same allele value does not mean there is a biological relationship; it just might be that many people would match as well.
Locus Alleged Father Child Percentage Index
D2S1338 12, 13 10, 12 1.845

However, it could be that the analysts already had results from many people reflecting allele values not of “12” at this location. That would make the uniqueness of the match stronger and more distinct. In such cases, the PI value could be greater from 3.452 to 12.819 or even 34.125. If there is no matching between profiles at the location, the PI value would be 0.000. When the profiles have finished being analyzed, the combined PI values for locations make up the Combined Parentage Index (CPI). For a father to be considered as the paternal or biological father of a child, the CPI number must be equal to or greater than 100. If the number is 100 or larger, then the probability of proof is at least 99.0000%, at 10,000 it increases to 99.9900%. A higher CPI means a closer percentage to 100%.

Locus Alleged Father Child Parentage Index
D2S1338

12, 13

10, 12

1.845

D2S1358

8, 11

11, 14

2.714

D8S1179

21.2, 32

19, 21.2

3.675

D19S433

15, 18

12, 15

8.338

Combined Parentage Index (CPI):

153.435

Probability of Paternity:

99.348%

To agree with guidelines for accreditation for DNA paternity tests, reports for paternity must reflect CPIs larger than 100 as well as a probability higher than 99.0% proving paternity. Accredited reports are required by most child support agencies, US courts, and other legal bodies (see Legal Paternity Testing for more information). For immigration cases CPI values of 200 (99.5%) or higher are required. Although results from the majority of paternal DNA tests follow this “text book” sample, there are too many other variables to not take them into consideration. For instance, some results can be skewed by mutations that occur naturally. On the other hand, even with many matches between father and child, it isn’t unheard of to receive a CPI value less than 100. However, in these cases if the maternal DNA is included, CPI values increase dramatically. Therefore we highly stress incorporating the biological mother when DNA testing for paternity. DNA Profiles A group of up to 16 markers makes up a standard DNA test profile applied to define identity and paternity. This profile also includes the FBI CODIS markers needed for DNA evidence. Genetic scientists created this profile for dependability in identifying each individual and named each marker in accordance with their location on the gene (like “D2S1338” is the 2nd gene with the location number “1338”). The location, or locus, of a marker has two alleles, denoted by integers on the report. Here alleles are just measures of short tandem repeats (STR) in the pattern of the gene. The two STR alleles at each locus come from both parents, one that is paternal one that is maternal.

DNA Profiles

Locus Alleged Father Child Parentage Index
D2S1338 12, 13 10, 12 1.845
D2S1358 8, 11 11, 14 2.714
D8S1179 21.2, 32 19, 21.2 3.675
D19S433 15, 18 12, 15 8.338

While analyzing a paternal DNA test, scientists seek to match values of allele numbers for the supposed father and child. For every marker, if the supposed father is the biological one, the child should have matches at every loci (natural mutations rarely occur, but when they do the child can have different values than the father. For the purposes of this explanation we assume there are no mutations). If there is just one mismatch between the father’s and child’s profiles, proof of paternity could be dismissed. Most often when proof of non-paternity is the result there are many mismatches. In the table below, the 3 mismatches provide proof of non-paternity. Although the supposed father is proven not to be biological in this instance, he still has 3 matches with the child. These matches however are not enough to provide proof of paternity. It is possible for any two people to have matches in their profiles, but a paternal correlation in alleles will reflect a match at every loci.

Locus Alleged Father Child Parentage Index
D2S1338

12, 13

8, 9

0.00

D2S1358

8, 11

13, 14

0.00

D8S1179

21.2, 32

19, 21.2

0.675

D10S1008

7, 12

15, 18

0.00

D14S1537

9, 14

11, 14

0.797

D19S433

15, 18

12, 15

1.338

Results Statement According to accreditation guidelines for paternal DNA testing, the outcomes must be stated in one of three possible ways. For instance, if John Smith is a supposed father, and James Smith the supposed child, and both were tested, the statement of paternity might be phrased one of the following ways: John Smith is excluded as the paternal or biological father of James Smith. In other words, it is not possible for John Smith to be the father of James Smith as the test did not reflect enough allele matches. John Smith is not excluded as the paternal or biological father of James Smith. In other words, the test shows that there is proof that Johns Smith is the paternal father of James Smith. You may ask, “Why does the report say “not excluded” if there is proof that he is?” Unfortunately, in DNA testing in order to give 100% proof of paternity the whole of the genome would have to be tested and matched between father and child. On the other hand, 100% proof of exclusion is easily done. Instead of testing the whole of the genome in a paternity test, the sample profile of 16 markers gives a probability of the relationship. These are sometimes probabilities in the order of 99.9999% positive that the alleged father has a biological relationship with the child.