The Genome Sequence of Drosophila melanogaster
Tóm tắt
Từ khóa
Tài liệu tham khảo
A number of methods were used to close gaps. Whenever possible gaps were localized to a chromosome region and a spanning genomic clone was identified. When a spanning clone could be identified it was used as a template for sequencing. The sequencing approach was determined by the gap size. For gaps smaller than 1 kb BAC templates were sequenced directly with custom primers. For gaps larger than 1 kb 3-kb plasmids or M13 clones from the clone-based draft sequencing were sequenced by directed methods or 10-kb plasmids from the WGS sequencing project were sequenced by random transposon-based methods. If no 3-kb or 10-kb plasmid could be identified PCR products were amplified from BAC clones or genomic DNA and end-sequenced directly with the PCR primers.
Henikoff S., Biochem. Biophys. Acta 1470, 1 (2000);
See ftp.ebi.ac.uk/pub/databases/edgp/sequence_sets/nuclear_cds_set.embl.v2.9.Z.
The genes found in unscaffolded sequence were Su(Ste) (FlyBase identifier FBgn0003582) on the Y chromosome His1 (FBgn0001195) and His4 (FBgn0001200) (histone genes were screened out before assembly) rbp13 (FBgn0014016) and idr (FBgn0020850).
M. G. Reese D. Kulp H. Tammana D. Haussler Genome Res. in press.
Sequence contigs were searched against publicly available sequence at the DNA level and as six-frame translations against public protein sequence data. DNA searches were against the invertebrate (INV) division of GenBank a set of 80 000 EST sequences produced at BDGP assembled to produce consensus sequences (21) and a set of curated Drosophila protein-coding genes prepared by three of the authors (M. Ashburner L. Bayraktaroglu and P. V. Benos) (15). Protein searches were performed against this set of curated protein sequences and against the nonredundant protein database available at the National Center for Biotechnology Information. Initial searches were performed with a version of BLAST2 (25) optimized for the Compaq Alpha architecture. Additional processing of each query-subject pair was performed to improve the alignments. All BLAST results having an expectation score of <1 × 10 −4 were then processed on the basis of their high-scoring pair (HSP) coordinates on the contig to remove redundant hits retaining hits that supported possible alternative splicing. This procedure was performed separately by hits to particular organisms so as not to exclude HSPs that support the same gene structure. Sequences producing BLAST hits judged to be informative nonredundant and sufficiently similar to the contig sequence were then realigned to the contig with Sim4 [
] for ESTs and with Lap [
] for proteins. Because both of these algorithms take splicing into account the resulting alignments usually respect intron-exon boundaries and thus facilitate human annotation. Some regions of the genome may be underannotated because the bulk of the annotation work was done on an earlier assembly version. Continued updates will be available through FlyBase.
M. G. Reese G. Hartzell N. L. Harris U. Ohler S. E. Lewis Genome Res. in press.
See the Gene Ontology Web site (www.geneontology.org).
See the Saccharomyces Genome Database Web site ().
D. Allen and J. Blake Mouse Genome Informatics (www.informatics.jax.org).
Hirano T., Curr. Opin. Genet. Dev. 10, 317 (1998);
C. Burge T. Tuschl P. Sharp in The RNA World R. Gesteland T. Cech J. Atkins Eds. (Cold Spring Harbor Laboratory Press Cold Spring Harbor NY ed. 2 1999).
See D. Nelson's Web site ().
High molecular weight genomic DNA was prepared from nuclei isolated [
] from 2.59 g of embryos of an isogenic y; cn bw sp strain [
]. The genomic DNA was randomly sheared end-polished with Bal31 nuclease/T4 DNA polymerase and carefully size-selected on 1% low-melting-point agarose. After ligation to BstX1 adaptors genomic fragments were inserted into BstX1-linearized plasmid vector. Libraries of 1.8 ± 0.2 kb were cloned in a high-copy pUC18 derivative and libraries of 9.8 ± 1.0 10.5 ± 1.0 and 11.5 ± 1.0 kbp were cloned in a medium-copy pBR322 derivative. High-throughput methods in 384-well format were implemented for plasmid growth alkaline lysis plasmid purification and ABI Big Dye Terminator DNA sequencing reactions. Sequence reads from the genomic libraries were generated over a 4-month period using 300 DNA analyzers (ABI Prism 3700). These reads represent more than 12× coverage of the 120-Mbp euchromatic portion of the Drosophila genome (Table 1). Base-calling was performed using 3700 Data Collection (PE Biosystems) and sequence data were transferred to a Unix computer environment for further processing. Error probabilities were assigned to each base with TraceTuner software developed at Paracel Inc. (www.paracel.com). The predicted error probability was used to trim each sequence read such that the overall accuracy of each trimmed read was predicted to be >98.5% and no single 50-bp region was less than 97% accurate. The efficacy of TraceTuner and the trimming algorithm was demonstrated by comparing trimmed sequence reads to high-quality finished sequence data from BDGP (Fig. 2).
For clone-based genomic sequencing BAC P1 and cosmid DNAs were prepared by alkaline lysis procedures and purified by CsCl gradient ultracentrifugation. DNA was randomly sheared and size-selected on LMP agarose for fragments in the 3-kb range for plasmids and in the 2-kb range for M13 clones. After blunt-ending with T4 DNA polymerase plasmids were generated by ligation to BstX1 adaptors and insertion into BstX1-linearized pOT2A vector. M13 clones were generated using the double-adaptor protocol [
]. Plasmid sequencing templates were prepared by alkaline lysis (Qiagen) or by PCR and M13 templates were prepared using the sodium perchlorate–glass fiber filter technique [
]. Paired end-sequences of 3-kb plasmid subclones were generated (principally) with ABI Big Dye Terminator chemistry on ABI 377 slab gel or ABI 3700 capillary sequencers. Additional M13 subclone sequence was generated using BODIPY dye-labeled primers. Procedures for finishing sequence to high quality at LBNL were as described (3).
J. F. Abril and R. Guigo Bioinformatics in press.
A. Peter et al. in preparation.
J. Locke L. Podemski N. Aippersbach H. Kemp R. Hodgetts in preparation.
The many participants from academic institutions are grateful for their various sources of support. We thank B. Thompson and his staff for the excellent laboratories and work environment M. Peterson and his team for computational support and V. Di Francesco S. Levy K. Chaturvedi D. Rusch C. Yan and V. Bonazzi for technical discussions and thoughtful advice. We are indebted to R. Guigo and to E. Lerner of Aquent Partners for assistance with illustrations. The work described was funded by Celera Genomics the Howard Hughes Medical Institute and NIH grant P50-HG00750 (G.M.R.).