ACCURATE BASECALLING AND QUALITY ESTIMATION FOR MODEL ORGANISM GENOME SEQUENCING

¹ Paracel, Inc. 3833 N. First Street San Jose, CA 95134

² Paracel, Inc. 80 S. Lake Avenue, Pasadena, CA 91101

The constant introduction of new technologies in sequencing and the ever-increasing volumes of genome sequence data from model organisms pose substantial challenges for software that base-calls, interprets and assembles reads from automated DNA sequencers. We have developed some sequence reconstruction software which is particularly suitable for low-pass or skim sequencing. A key component of this is obtaining highly accurate base calls and estimates of sequence quality from automated sequencers, for which we have developed some novel algorithms. Our software, which initially is supporting sequencing projects for the rice, drosophila and arabidopsis genomes using the ABI 3700 instrument, adjusts the base calls from the sequencer and assigns an estimate of error probability which is calibrated specifically for the instruments and conditions used in those projects. We present results that indicate that our algorithms significantly extend read lengths, base calling accuracy and error estimate accuracy when compared to existing tools. This higher resolution data then leads to better results with assembly tools that are optimized for low pass sequencing.