An expert system for scoring DNA database profiles

摘要: Forensic DNA databases are becoming an increasingly valuable law enforcement tool for convicting repeat offenders and exonerating the innocent. However, constructing such is quite laborious. After generating STR profiles in lab, people expend even greater effort visually reviewing data before it enters database. All artifacts must be detected, no error can tolerated. With millions of forensic samples to analyzed, this bottleneck has become a formidable task. We have developed software analysis methods that automate review potentially eliminate 90% work. Our fully automated TrueAlleleTM system inputs raw fluorescent sequencer (gel or capillary) files, processes gel image (separating colors, tracking sizing lanes), analyzes experiments (quantitating peaks, comparing with ladder calling alleles). For each allele call, assigns quality score applies artifact detection rules. These checks enable user focus on just 10% suspect data, thereby eliminating most effort. models every step generation process. By computing hundreds variables genotype, compare observed against expected behavior. Large deviations between flag problematic human review. automatically extracts information needs from (e.g., ABI/377 collection files). processing, provides assurance assessing runs (lane tracking, quality, control lanes, etc.). Automated processing continues peak quantitation, designation, scoring. In allele-based review, focuses those designations which flagged as having specific problems. (and possibly editing) small subset then generates files submission database CODIS). operates independently manufacturer technology, all major computer platforms (Macintosh, Windows, UNIX). The program analyze any panel loci, allelic ladders, internal size standards. Evaluation available “www.cybgen.com”. expert designed provide computer-based “second scorer” profiles. British Science Service (FSS) selected scoring scaling up UK National Database. anticipate will role significantly reducing providing laboratory feedback, automating