DNA Identification in Mass Fatality Incidents
Robert Shaler
The Office of the Chief Medical Examiner (OCME) partnered with other laboratories in the World Trade Center identification effort. The OCME was in the management role—with ultimate responsibility for making the victim identifications—and also performed retesting of the remains and secondary testing of family samples and personal effects. However, the primary testing of bones and tissues and the initial testing of family and reference samples was contracted to outside laboratories with specialized experience.
A centralized project structure, where all samples are accessioned and analyzed by a single laboratory, is the most common paradigm for sample receipt and analysis. The term “centralized” does not necessarily imply a specific physical location or software/hardware architecture.
In a decentralized project structure, more than one laboratory is involved. The laboratory that is ultimately responsible for the mass fatality incident response is called the managing laboratory, and other participating laboratories are referred to as partner laboratories.
Since a mass fatality DNA identification effort most likely would be added to a laboratory’s casework, a decentralized structure can be more efficient if good information technology support exists. For example, in a decentralized structure, the human remains samples might be analyzed by one laboratory and the reference samples by another. Both laboratories would analyze samples independently, leveraging their respective strengths, and the overall response undoubtedly would be faster because the laboratories would be working simultaneously.
However, for a decentralized model to work well, there must be a mechanism for centralized data management so that the managing laboratory and the partner laboratory/laboratories can view information and communicate about data, regardless of where they are collected or analyzed. It is especially critical that the managing laboratory have as close to real-time access as possible to all data—including DNA profiles, chain-of-custody documentation, and metadata—that is associated with the mass fatality incident response, because the managing laboratory has the ultimate responsibility for making comprehensive and frequent updates to the families, public officials, the media, and the public. [Note: For example, metadata for a victim’s toothbrush include the name of the victim, and when and by whom it was provided.]
Sample accessioning and sample storage can be decentralized as long as each partner laboratory ensures that all of the metadata are accessible by the managing laboratory. The physical samples can be stored at partner laboratories, as well. However, if a partner laboratory disengages from the response effort, all of its samples must be shipped to the managing laboratory under appropriate chain-of-custody procedures.
It is important to consider the administrative review portion of the identification process when deciding between a centralized or decentralized project structure. In a decentralized plan, if the managing laboratory needs to examine the physical item (e.g., toothbrush, hairbrush), the partner laboratory must be prepared to pull the physical item from evidence storage and send it to the managing laboratory or be prepared to provide digital images that can be accessed electronically.
Sample analysis readily lends itself to a decentralized structure. A managing laboratory can divide the workload in several different ways: by DNA technology, by sample type, or some by combination thereof. For example, bone fragments might be shipped to one laboratory for STR and mtDNA analyses, whereas tissue samples would be analyzed in-house. Or partner laboratories might receive entire samples or extracted DNA from the managing laboratory.
Because the managing laboratory has the ultimate responsibility for maintaining the chain of custody for samples, extracts, and data, it is important to recognize the management challenges presented by these aspects of a mass fatality identification effort. Samples, extracts, and data may be shipped to and from the managing laboratory individually or collectively, at different times and in different batches. Multiple extracts and multiple DNA profiles (data) might be derived from a single sample, and the laboratory’s sample tracking system must be able to document and certify the chain of custody for each one. The sample tracking system must collect data associated with all physical transfers, including what was sent, where it was sent, when it was sent, when it was received, and by whom. The managing laboratory uses this information to document the chain of custody and to provide status updates to the public.
In a decentralized structure with multiple partner laboratories, the managing laboratory must decide how samples, extracts, and data will move among the partners. There are two basic approaches: the daisy-chain model and the hub-and-spoke model. Both were used in different aspects of the World Trade Center (WTC) DNA identification project.
The Daisy-Chain Management Model
In the daisy-chain model, samples (or extracts) are shipped to the first partner laboratory, which ships extracts to the second partner laboratory, which ships extracts to the third partner laboratory, and so on. A major drawback to daisy-chaining is that the managing laboratory does not physically control the flow of samples and extracts among partner laboratories, which can make it difficult to locate missing samples and ensure proper chain-of- custody documentation.
When the managing laboratory has a partner laboratory perform DNA extraction, the daisy-chain model can become very convoluted. Exhibit 8 depicts the flow of samples, extracts, and data in a modified daisy-chain structure that would occur when the managing laboratory sends samples to the first partner laboratory for extraction and analysis. In this example, the first laboratory extracts a sufficient quantity of DNA for the two other partner laboratories and ships the extracts to them. Partner laboratories return leftover extracts and data to the managing laboratory. It is important to note that the daisy-chain model requires compatible informatics and hardware systems and shared data transfer protocols so that all parties are sharing information.
The Hub-and-Spoke Management Model
In the hub-and-spoke model (see exhibit 9), the managing laboratory centralizes the control and movement of samples, extracts, and data among partner laboratories. Although it is a simpler model than the daisy-chain for tracking chain of custody, locating missing samples, and identifying missing data, there are some limitations. The major disadvantage of a hub-and-spoke structure (in addition to time delays) is that samples or extracts must be packaged and shipped multiple times, which could result in repeated freezing and thawing, potentially decreasing the quality of the DNA in the samples. For example, the managing laboratory sends a batch of samples to the partner laboratory for extraction. When the managing laboratory receives the extracts back from the partner laboratory, it must open the package and verify the contents, then repackage the contents and ship them to the second partner laboratory. This doubles the work as compared with a daisy-chain structure. However, in a hub-and-spoke structure, the managing laboratory has a higher level of control and the possibility of miscommunication between partner laboratories is reduced.
Because of the types of samples in a mass fatality incident (e.g., bone fragments, tissue, personal items, kinship swabs) and the numerous DNA technologies (short tandem repeats, mitochondrial DNA, single nucleotide polymorphisms, etc.), a decentralized structure is often necessary. Moreover, it may be prudent to create different workflow mechanisms for different types of samples. For example, kinship samples may be processed using a daisy-chain model, whereas disaster samples may be better handled using a hub-and-spoke system. Regardless of which project structure is used, however, it is safe to assume that the greater the number of partners, the greater the management complexity.
Identification management is one function that should never be decentralized. The managing laboratory is responsible for setting the parameters for DNA identifications and resolving conflicts with other identification modalities. The managing laboratory also acts as the single point of contact for the victims’ families, public officials, and the media on identification-related matters. Thus, it is critical that all data—metadata and DNA profiles—be provided to the managing laboratory.
