DNA Profiling, Sequencing, and Genomics

DNA profiling (DNA fingerprinting) identifies individuals by comparing highly variable regions of their DNA, DNA sequencing determines the exact order of the bases in a DNA molecule, and genomics is the study of whole genomes. In brief: forensic DNA profiles are based on short tandem repeats (STRs), short sequences repeated a variable number of times, which differ between people. Matching profiles at 13 to 20 STR loci makes the chance that two unrelated people have the same profile extremely small. The human genome has about 3.2 billion base pairs, and sequencing it cost about 3 billion dollars (the Human Genome Project), whereas today the cost of sequencing a genome is on the order of a few hundred dollars.

These techniques have changed forensic science, medicine, ancestry research and conservation. This guide explains how DNA profiling works, what STRs are, how the match probability is calculated, outlines Sanger and next-generation sequencing, introduces genome analysis, works through a profile-matching and probability example and corrects common mistakes.

What is DNA profiling?

About 99.9 percent of the DNA sequence is the same in all humans. The remaining 0.1 percent includes regions that differ a lot between individuals. The most useful for identification are short tandem repeats (STRs): short sequences of 2 to 6 bases (for example, GATA) that are repeated one after the other, with the number of repeats differing from person to person. A person inherits one allele (a number of repeats) from each parent at each STR locus. For example, at one locus, a person might have 8 repeats from the mother and 11 repeats from the father, which is written as the genotype 8,11.

The repeat numbers at many loci give a unique combination in everyone except identical twins. In the United States, the FBI’s CODIS system uses 20 core STR loci, and in Europe a standard set of 12 to 16 loci is used. Related individuals share alleles, so profiles can also be used to determine paternity and other family relations.

What are the steps of DNA profiling?

1. Collect and extract DNA from blood, saliva, hair roots, skin cells or other tissue. Only a tiny amount is needed (less than a nanogram).
2. Amplify the STR loci by PCR with fluorescently labeled primers that flank each STR. This makes enough DNA and tags it.
3. Separate the fragments by size using capillary electrophoresis, a version of gel electrophoresis in thin tubes. A detector reads the fluorescence, and fragment lengths are calculated automatically.
4. Compare the profile (a series of peaks showing the alleles at each locus) with others, such as from a crime scene and a suspect, or a database.
5. Calculate the probability of a random match using the frequencies of the alleles in the population.

Mitochondrial DNA is used for degraded samples such as old bones, as there are many copies per cell, and it is inherited through the mother only. Y-chromosome STRs trace the male line.

How is DNA sequenced?

Sanger (chain-termination) sequencing was the main method for about 30 years. DNA is copied by DNA polymerase in the presence of normal nucleotides and a small amount of modified nucleotides (dideoxynucleotides, ddNTPs), each labeled with a different fluorescent color. When a ddNTP is added, the chain stops, because it lacks the 3′ hydroxyl group needed to attach the next nucleotide. The result is a set of fragments of all possible lengths, each ending with a labeled base. They are separated by size in capillary electrophoresis, and the sequence is read from the colors, in order of size. It is accurate for reads of up to about 800 to 1 000 bases.

Next-generation sequencing (NGS) reads millions or billions of short fragments (typically 50 to 300 bases) in parallel. DNA is fragmented, attached to a surface, copied and read by detecting the nucleotides added to each fragment, usually with fluorescent signals. Computers assemble the sequence by matching overlapping fragments to each other or to a reference genome. A genome is sequenced many times (e.g. 30× coverage) for accuracy. Third-generation (long-read) sequencing, such as nanopore sequencing, reads single DNA molecules of thousands to millions of bases, as the DNA passes through a tiny pore and changes the electric current, which helps to resolve repetitive regions. These improvements lowered the cost per genome from about 100 million dollars in 2001 to a few hundred dollars today.

What is genome analysis?

Genomics studies the complete set of DNA in an organism. After sequencing, bioinformatics software is used for:

• Annotation: finding genes and regulatory regions. The human genome contains about 20 000 protein-coding genes, which make up only about 1 to 2 percent of the DNA.
• Comparison: aligning genomes of different individuals or species to find variants (such as single nucleotide polymorphisms, SNPs) and evolutionary relationships. Humans and chimpanzees differ by about 1 to 2 percent in the aligned regions.
• Medical genomics: finding mutations that cause or raise the risk of disease, such as the BRCA1 gene for breast cancer risk, and selecting treatments (pharmacogenomics).
• Pathogen genomics: tracking the spread and evolution of viruses and bacteria, as in COVID-19 variants.

Comparison of the methods

Worked example: matching a profile and calculating probability

DNA from a crime scene is compared with a suspect at four STR loci. The numbers are the repeat counts of the two alleles at each locus.

1. Suspect A: matches at all four loci. Suspect B: differs at locus 2 (15 instead of 16), so Suspect B is excluded. One clear mismatch is enough to exclude a person, unless there is an error in the test.
2. Probability of a chance match for Suspect A: use the product rule. The genotype frequency for a heterozygote with allele frequencies p and q is 2pq. Suppose the frequencies are: locus 1, alleles 8 and 11: 0.10 and 0.08; locus 2, alleles 14 and 16: 0.12 and 0.05; locus 3, alleles 9 and 12: 0.09 and 0.11; locus 4, alleles 6 and 7: 0.07 and 0.06.
3. Genotype frequencies: locus 1: 2 × 0.10 × 0.08 = 0.016; locus 2: 2 × 0.12 × 0.05 = 0.012; locus 3: 2 × 0.09 × 0.11 = 0.0198; locus 4: 2 × 0.07 × 0.06 = 0.0084.
4. Combined (the loci are independent): 0.016 × 0.012 × 0.0198 × 0.0084 = 3.19 × 10⁻⁸, i.e. about 1 in 31 million.

With only four loci, about 1 in 31 million, a match is already strong, and 20 loci give probabilities of the order of one in a quintillion or smaller. This number is not the probability that the suspect is guilty. It is the probability that a random unrelated person would match. Relatives, laboratory error, contamination or mixed samples can also complicate interpretations, and DNA evidence must be considered together with other evidence, which is known as avoiding the “prosecutor’s fallacy”.

What are the applications and issues?

• Forensics: identifying suspects and victims, exonerating people who were wrongly convicted (the Innocence Project has used DNA tests in hundreds of cases). Investigative genetic genealogy uses consumer ancestry databases to find relatives of an unknown suspect, and it solved cases such as the Golden State Killer in 2018.
• Paternity and kinship tests; immigration cases.
• Medicine: diagnosing rare diseases by exome or genome sequencing, screening newborns, tumor sequencing to choose therapy.
• Conservation and agriculture: identifying species and tracking illegal wildlife trade, breeding programs.
• Issues: privacy and consent, the storage of DNA databases, genetic discrimination, the possibility of finding unexpected family information, and fairness in the access to genomic medicine.

How did these methods develop?

Alec Jeffreys developed DNA fingerprinting at the University of Leicester in 1984 using minisatellite probes, and it was first used in a criminal case in 1986 to 1987, in the Colin Pitchfork murder case in England, in which it exonerated an innocent suspect and later identified the culprit. Frederick Sanger developed his sequencing method in 1977 and received his second Nobel Prize in Chemistry in 1980. The Human Genome Project ran from 1990 to 2003, and the first draft was published in 2001. NGS appeared in the mid-2000s. The first complete, gap-free human genome sequence was reported in 2022.

Which mistakes are common?

Mistake 1: Saying that DNA profiling compares whole genomes

Forensic profiling compares a small number of variable STR loci, not the full sequence.

Mistake 2: Misreading the match probability

It is the chance that a random person has the same profile. It is not the probability that the suspect is guilty.

Mistake 3: Confusing profiling with sequencing

Profiling counts repeats at specific places. Sequencing reads every base.

Mistake 4: Multiplying probabilities for related loci

The product rule requires independent loci. Relatives and linked loci need special treatment.