Characterizing patterns in DNA sequence trace data through informatics tools

Abstract

Sequence trace data files produced by the Sanger sequencing method were believed to create peak height values of random height and with no added value for base calling. Our study is the first to comprehensively prove the existence of definable peak height patterns and to develop tools that allow the characterization of the frequency of such patterns for each sequence frame. By studying hundreds of mtDNA samples sequenced in two certified forensic laboratories, in the United States of America and in Portugal, we were able to prove that peak height patterns are predictable and the same from sample to sample if the chemistry and primer combination is kept constant within the same laboratory. Moreover, the characterization of these patterns and the ability to predict their behavior for other samples led us to develop the novel concept of Sequence Biometrics. Sequence Biometrics defines the characteristics of these peak height patterns for a certain stretch of sequenced DNA, independently of the origin and sample, which is specific to the primer/chemistry combination used within the laboratory. Therefore, Sequence Biometrics is a new quality parameter for sample processing and can be used by novel expert systems in the assessment of new data. This work provides the basic informatics tools and workflow mechanisms to build standard Sequence Biometrics tables, of several primer/chemistry combinations and to define their characteristics and boundaries of its use.