Sequencing Ustilago maydis with the
LI-COR 4200 DNA Sequencer

Dr. Dagmar Schuette, Sequencing Department
LION Bioscience AG, Heidelberg, Germany

Introduction

In contrast to most microbial genome projects, yet similar to other eucaryotic genome projects like S. cerevisiae or S. pombe, C. elegans and D. melanogaster, it was decided to analyze the Ustilago maydis genome in an ordered manner and not by a whole genome shotgun approach. A chromosome-specific minimal tiling path of overlapping BAC clones, covering approximately 90% of the genome, was established to provide sequence-ready maps of the genome. Individual BAC clones contributing to the minimal tiling path were further subcloned into sets of minimal overlapping plasmid subclones, which then served as sequencing templates in a "single pass" sequencing approach. Therefore, we expected to reach a coverage lower than 3x at the end of the project.

Under these circumstances, it was necessary to use a sequencing system that provided long read lengths combined with high accuracy because we expected large single stranded regions.

For this purpose, the technology of the LI-COR sequencing machines appeared to be the right choice, not only because of their long read length and high accuracy, but also because of the robust system, easy handling and the sequencer's relative insensitivity to contamination due to the infrared detection system.

Based on average read lengths of 800-1000 base pairs provided by the LI-COR machines, plasmid subclones were constructed at an average size of 1.6-1.7 kb. Considering the insert size and number of base pairs received from the forward and reverse read, we expected complete sequencing of the inserts in one run with an overlap of good quality.

Physical maps from the subclones of each BAC were generated (Fig. 1) and used to construct a minimal tiling path, expecting an average overlap between the clones of up to 200 base pairs. The clones selected from this map served as sequencing templates. Due to biological or technical problems, the sequence analysis led to 5-10 contigs per BAC. Gap closure was performed in a two-step process. First, sequencing alternative clones were selected for that issue from the physical map. Remaining gaps were closed by primerwalks on BACs on the ABI sequencer 377 (Fig. 2).

Fig. 1. Physical map of subclones generated from a BAC. Dig-labeled probes have been hybridized against clones, bound on filters. The clone order results from the hybridization signals. Fig. 2. The order of the clones belonging to the minimal tiling path is verified by the order after assembly.

We have sequenced 17,425,931 base pairs (bp) of the Ustilago maydis genome with very high accuracy out of 28 large contigs comprising 258 minimal overlapping BAC clones covering the 23 chromosomes of the organism. Only 22 small sequencing gaps remained open. Compared to the estimated genome size of 20.5 megabases, we have covered approximately 85.9% of the genome, leaving mainly ribosomal repeat areas, telomeric regions and other unclonable regions untouched.