ESTIMATING, ENCODING, AND UTILIZING UNCERTAINTIES IN SEQUENCE DATA

AMONG THE STATED GOALS OF THE HUMAN GENOME PROJECT IS A DRAMATIC IMPROVEMENT IN DNA SEQUENCING TECHNOLOGIES AND CORRESPONDING REDUCTION IN COST PER FINISHED BASE SEQUENCED. RELATIVELY LITTLE ATTENTION, HOWEVER, HAS BEEN GIVEN TO THE PROBLEM OF ASSURING THE QUALITY OF THE RISING TIDE OF MASS-PRODUCED DATA AS GENOMIC SEQUENCING BECOMES EVER MORE AUTOMATED AND PRODUCTION-ORIENTED. AT A VERY MINIMUM, IT WILL BE INCUMBENT ON THE PRODUCERS OF SEQUENCE DATA TO PROVIDE INFORMATION REGARDING ITS RELIABILITY, PREFERABLY FOR EACH SEQUENCED BASE. THIS WILL REQUIRE THAT SUCH INFORMATION BE ESTIMATED AS AUTOMATICALLY AND AS INEXPENSIVELY AS IS THE SEQUENCE DATA ITSELF. THIS PHASE I PROJECT WILL ADDRESS THIS ISSUE BY ACHIEVING THE FOLLOWING BROAD OBJECTIVES: (1) APPROACHES TO THE ESTIMATION OF BASE-SPECIFIC UNCERTAINTY INFORMATION FROM THE OUTPUTS OF AUTOMATED SEQUENCING MACHINES WILL BE STUDIES AND EVALUATED, (2) METHODS FOR COMPACTLY ENCODING, STORING, VIEWING, AND MANIPULATING THIS INFORMATION WILL BE IMPLEMENTED AND EVALUATED, (3) A FRAGMENT ASSEMBLY ALGORITHM WILL BE DESIGNED THAT UTILIZES THIS INFORMATION FOR INPUT FRAGMENTS AND GENERATES CONSISTENT ESTIMATES FOR FINISHED ASSEMBLIES, AND (4) A VARIETY OF SEQUENCE ANALYSIS ALGORITHMS WILL BE ADAPTED TO MAKE OPTIMAL USE OF UNCERTAINTY INFORMATION WHERE IT IS AVAILABLE.