Automated DNA extraction from samples eases the burden on forensic laboratories and results in more consistent processing. A number of kits specifically developed for use with these extraction techniques are expected to drastically reduce the turnaround time for forensic cases.

Since Alec Jeffreys and his colleagues published their seminal papers on minisatellite regions in human DNA^1-3 in 1985, techniques based on DNA fingerprinting have become well established for solving problems of human identification. Each individual’s polymorphic variation within those minisatellites was found to be sufficiently distinctive that scientists could estimate the likelihood that a biological sample either came from, or was related to, that person. As well as the obvious crime scene identification, tissue remains such as those from soldiers killed in action (see below), can be used to identify the individual. In other legal cases, DNA samples can be compared to prove or disprove maternity,paternity, or other family relationships.

By the late 1980s, DNA profiling was being used around the world and, with the arrival of PCR methods,^4,5 DNA fingerprinting from much smaller samples, including single cells, became possible. This development, together with the identification of microsatellites,⁶ led to the possibility for DNA profiling to be automated. Following the DNA Identification Act of 1994, the FBI was formally able to establish a national criminal database and the National DNA Index System (NDIS) was established in 1998, as part of the Combined DNA Index System (CODIS). In December 2005, 50 American States were actively participating in NDIS, which included almost three million differentoffender and forensic sample profiles.

DNA can now be extracted from a wide range of samples, such as blood, semen, hair roots, bones, or saliva, and including stains or skeletal remains that are many years old. Since the first American case in 1987, when Tommie Lee Andrews was found guilty of rape and burglary,⁷ numerous successful court convictions have taken place where evidence has included the matching of DNA obtained from a crime scene with the defendant’s DNA profile. However, the availabilityof biological samples left behind at a crime scene is usually very limited ä andbeing able to carry out DNA profiling from such minute sources relies both onthe successful extraction of DNA as well as its subsequent characterization and analysis. Furthermore, a number of problems exist with DNA typing of samples such as stains on materials or old, degraded samples, including contamination or poor quality of the DNA obtained.^8,9 As a result, purification of DNA from samples is still a rate-limiting step in obtaining useful genotypes. Continuing improvements in techniques for extracting DNA from forensic samples will allow more cases to be investigated using DNA profiling, as well as improve the reliabilityof the method and speed up sample processing time.

More recently, significant technological advances have been made that are changing DNA testing laboratories around the world. Early methods of automated DNA extraction began in the late 1980s and laboratory automation instruments with extremely high throughput and reliability are now widely available. Today’s liquid handling workstations are capable of processing a great variety of complex protocols and come in a range of sizes and capacities, making them suitable for all types of forensic laboratories. The simultaneous introduction of off-the-shelf kits specifically designed for DNA isolation from various sample types has equally set new expectations of the quality, sensitivity, and consistency of extracted DNA for forensicanalysis, and some of these recent innovations are discussed below.

Automating Forensic Sample Processing
The labor-intensive workflow involved in processing forensic samples typically includes sample collection, DNA isolation, DNA quantitation using Real-Time PCR, analysis of short tandem repeat (STR) loci with PCR and analysis of the amplified STR sizes, analysis of the DNA profile and comparison to known samples, and/or submission of the profile to database(s) in search of matches. Of these steps, it is currently possible to fully automate the DNA isolation, quantitation, amplification, and STR preparation/reaction steps. Such automation can massively reduce labor and man-hours within the laboratory as well as minimize handling error and improve processing consistency, allowing reproducible purification of small amounts of DNA from a wide range of samples. Moreover, laboratory personnel can be freed up for involvement in other steps, such as profile analysis, therefore significantly reducing casework turnaround time. The use of automated liquid handling workstations has become increasingly widespread through forensics laboratories worldwide over the last decade. These set-ups were initially of particular benefit for the so-called ‘databasing’ laboratories involved in processing huge numbers of reference and/or offenders’ samplesfor database information.

For example, researchers at the Armed Forces DNA Identification Laboratory (AFDIL) in Maryland have drastically increased their sample throughput by developing automated methods for processing mitochondrial DNA from skeletal remains. Applying DNA typing techniques to successfully establish the identity of a decomposed homicide victim was first published in 1990¹⁰ and, since then, methods have been developed to allow identification of human remains that are decades old. However, standard DNA fingerprinting can only be achieved if relatively high molecular weight DNA is extracted from the tissue samples and, after a long post-mortemdelay, many tissues such as blood and kidney are too degraded to be used.¹¹

Tim McMahon and his colleagues at AFDIL are involved in identifying the remains of American service members killed in action, from recent conflicts including Iraq or older campaigns such as Korea, Vietnam, and World War II. While the more recent casualties can be identified from fresh bone and tissue samples, the older skeletal remains present more of a problem as their nuclear genomes are often too degraded for standard nuclear DNA (nucDNA) identification methods. Instead, identification may be achieved by sequencing mito-chondrial DNA (mtDNA),¹² which is present in higher concentrations in degraded remains compared to genomic DNA. mtDNA can be extracted and the hypervariable regions one and two (HV1 and HV2) amplified using primer sets or mini primer sets, depending on the level of degradation. The results of these amplifications are then compared to reference samples collected from maternal relatives of missing forces personnel and, together with historical records, anthropological and after-action battlereports are used to propose a formal identification.

AFDIL receives over 800 bone samples from unidentified bodies and between 1,500 and 2,000 family reference samples per year. The bone samples, because of their general condition, require manual DNA extraction and sequencing but the reference samples are more straightforward to process and are sequenced using seven primers that cover over 1,122 bases of the mtDNA control region that contains HV1 and HV2 as well as mini variable regions 1 and 2. Samples are processed in groups of 93 plus two negative controls and a PCR positive control. Preparing individual sequencing reaction plates for each primer, placing them individually into the thermal cyclers, and purifying the samples after sequencing is therefore manually intensive, equating to as much as ¹² man-hours. However, with an automated liquid handling workstation,AFDIL has developed a faster method of preparing the sequencing reactions that also offers improved sensitivity and low contamination levels. The completely automated set-up allows the user to walk away once the instrument is running and return the following morning when sequencing is complete. This hands-free process has resulted in a significant drop in the number of samples that have had to be re-sequenced. Previously, on average, this would equate to 4.9 percent but, using the workstation, the failure rate has fallen by a factor of ten, saving time, money, and reagents. Similarly, the time taken to process the samples has been cut by almost two thirds, giving technicians more time to concentrate on sequencing the bone samples and for data analysis. Automation of the reference sample sequencing process has undoubtedly accelerated AFDIL’s casework processing, helping to answer some decade-old questions of identityand bring peace to the families of soldiers lost in conflict.