Draft Genome Sequence of Geotrichum candidum Strain 3C

Abstract

We report here the draft genome sequence of Geotrichum candidum strain 3C, which is a filamentous yeast-like fungus that holds great promise for biotechnology. The genome was sequenced using Ion Torrent and 454 platforms. The estimated genome size was 41.4 Mb, and 14,579 protein-coding genes were predicted ab initio.

GENOME ANNOUNCEMENT

Geotrichum candidum is a filamentous yeast-like fungus. The cellulolytic enzymes from G. candidum attract the most interest due to their potential applications in many fields such as protoplast generation and the textile, paper, and detergent industries (1, 2). In recent years, growing attention has been drawn to fuel ethanol production by using cellulases (3). Conversely, G. candidum and related species may degrade various natural and artificial materials. A species close to G. candidum was found to be responsible for biodeterioration in compact discs (CDs), destroying the information pits (4). The G. candidum strain 3C itself was found and isolated from a rotting rope (5). Previous studies demonstrated that this strain possesses high cellulolytic and xylanolytic activities (6, 7). So, this fungus holds great promise for biotechnology; however, there are no genome sequences of G. candidum strains available at the present time.

We sequenced the whole genome of G. candidum strain C3 using Roche 454 and Ion Torrent (Life Technologies) platforms. The DNA library for Roche 454 GS Junior was prepared using GS Rapid Library Prep Kit and sequenced with the GS Junior Titanium sequencing kit. The amount of raw data was more than 84 Mb in 182,317 reads with an average read length of 460 bp. This appeared to be insufficient for the de novo genome assembly, and we sequenced the genomic DNA using Ion Torrent PGM with an Ion Plus fragment library kit, an Ion PGM sequencing 400 kit, and Ion 318 Chip v2 to get additional information. With the Ion PGM instrument we obtained 1.3-G total bases in 4.39 million reads with a median read length of 343 bp. Both datasets were used for the de novo genome assembly using MIRA 4.0 (8). Genome assembly resulted in 560 large contigs (500 bp or more) with a total consensus length of 41,384,521 bp, largest contig size of 1,363,582 bp, and N₅₀ of 437,602 bp. The average total coverage was 31.25, and the G+C content was 46.96%. We masked the repetitive sequences using RepeatMasker version open-4.0.5 (http://www.repeatmasker.org), and used the self-training gene-prediction software GeneMark-ES 2.0 (9) to create a training data set for Augustus. We used Augustus 3.0.2 (10) trained with the output data from GeneMark-ES for ab initio gene prediction. In this way we could identify 14,579 protein-coding genes.

Nucleotide sequence accession numbers.

This whole-genome shotgun project has been deposited at DDBJ/EMBL/GenBank under the accession no. JMRO00000000. The version described in this paper is version JMRO01000000.