The following evidence types were aligned to BAC contigs using Exonerate (Slater GS, Birney E. Automated generation of heuristics for biological sequence comparison. BMC Bioinformatics. 2005 Feb 15;6:31).

• 36,430 Ceres full-length cDNAs [retrieved 2008 Nov 7]
• 14,963 Arizona full-length cDNAs [retrieved 2008 Nov 7]
• 1,462,607+ 537,726 ESTs and another 18181 mRNAs from maize [retrieved 2008 Nov 7]
• 1,217,859 ESTs and another 72,919 mRNA from rice [retrieved 2008 Nov 7]
• 2,448,641 ESTs and another 14015 mRNAs from other monocots [retrieved 2008 Nov 7]
• 359,942 SwissProt proteins from all species [retrieved 2008 Jul 7]
• 494,444 non-maize plant proteins in Trembl [retrieved 2008 Jul 7]
• 94,734 GenBank proteins from plant species [retrieved 2008 Aug 20]
• 52,177 rice proteins from rice gene annotations [retrieved 2008 Aug 20]
• 36,338 proteins from sorghum gene annotations [retrieved 2008 Aug 20]

Alignment data were integrated into gene models using a computational pipeline based on Ensembl GeneBuilder (Curwen V, Eyras E, Andrews TD, Clarke L, Mongin E, Searle SM, Clamp M. The Ensembl automatic gene annotation system. Genome Res. 2004 May;14(5):942-50).

The Working Gene Set provides an overall representation of non-overlapping candidate gene elements across the maize genome sequence. It is constructed by complementing evidence-based genes predicted by GeneBuilder with ab initio gene models predicted by Fgenesh on RepeatMasked sequence. Only Fgenesh models that do not overlap with the evidence-based genes are used in the set.

The Working Gene Set can be accessed on the site in a number of ways. The release page provides rudimentary statistics about the working set. Clone-oriented (ContigView) displays show the working set as a track, where each gene type is distinguished by color. Finally, the gene set itself can be downloaded through the FTP site

We have transformed the FPC bands to genomic coordinates using a conversion factor of 4900 bases per band. This number, however, should be treated as an approximation.

In a 2005 paper, Ed Coe and Mary Schaeffer address these questions (page 299):

For the U.S. public genomic project, BAC libraries were initiated in 1998-1999. Three BAC libraries were constructed from the B73 inbred line, one of the two IBM mapping parents, for use in generating a physical map. The specific germplasm source for the B73 inbred line, PI 550473, is continuously and reliably available from the USDA North Central Regional PI Station at Ames, Iowa, and can be ordered through the Stock link at MaizeGDB.

Three different restriction enzymes were used to construct the libraries, to ensure comprehensive genome coverage. The HindIII library of 247,680 clones, ZMMBBb, constructed at Clemson University Genomics Institute (CUGI) (2005), has an average insert size of 137 kB and a genome coverage of ~17X. Characterizations showed this library to be representative and to be virtually free of chloroplast sequences (Tomkins et al, 2002). The HindIII library is available from the Arizona Genomics Institute (AGI) or from CUGI. An EcoRI BAC library, ZMMBBc (CHORI 201 segment 1), constructed at Children's Hospital Oakland Research Institute (CHORI), has an average insert size of 163 kB and a genome coverage of ~6.9X. An MboI library, ZMMBBc (CHORI 201 segment 2), constructed at CHORI in collaboration with Jo Messing, has an average insert size of 167 kB and a genome converage of ~7X. The EcoRI and MboI libraries are described by and can be ordered from CHORI. Hybridization filters for all three libraries are available from CUGI, AGI, and CHORI.

The paper provides further information about the BAC construction strategy. (made available with permission from Maydica)

Since the project aims to sequence the maize genespace, it is expected that regions on BACs that do not contain genes will not be sequenced to a finished quality. For more information, read about the various maize sequencing statuses.

Yes. The menu above the browser window lets you control which tracks are visible and which are not. Furthermore, the browser uses cookies to remember the decision for 2 months.

Yes. The Ensembl browser supports DAS (Distributed Annotation System). Simply configure a URL to your custom data source (either through a pull-down menu or from the sidebar) and a dedicated track should appear. We now provide pre-configured tracks for maize TwinScan transcripts produced by the Danforth Center.

Virtual bins are computational representations of genetic bins found on the maize genetic maps. We developed virtual bins by generating correspondences between the AGI FPC map and the IBM2 Neighbors map. Markers common to both maps on the same chromosome were then assigned chromosomal positions based on the AGI approximation of 4900 basepairs per fingerprint contig band and then clustered with respect to maize bin. For any given bin, those markers outside of three deviations from the mean were discounted as spurious. Virtual bin boundaries were automatically calculated from the remaining minimum and maximum positions of markers in a cluster. Following automatic generation, we manually curated bin positions to assure no overlaps.

The mathematically-defined repeat (MDR) track on ContigView is a density plot describing how repetitive a stretch of DNA is. The frequency of each constituent 20-mer along the BAC sequence was determined within the raw reads of the 0.45X maize whole genome shotgun sequence (DOE Joint Genome Institute). The track is presented in a logarithmic scale, since the distribution of oligomer copies varies widely across the genome (from 1 to more than 10,000 copies).

Predicted peptides are screened to distinguish those encoded by transposable elements (TE). Sequences are aligned to the NCBI non-redundant protein database using BLASTP. Those that align to transposable element sequences, as specified within a curated database, are classified as TE and are color-coded maroon within the browser's ContigView. Those not classified as TE are color-coded blue. FGENESH can predict coding sequences that incorrectly contain components of two genes. Approximately 3% of genes classified as TE are "fusions" that contain portions of non-TE genes.

Prediction translations are run through NCBI BLASTP using the default parameters. By default, the BLAST low-complexity filter is on. The website, on the other hand, turns off the filter by default, showing alignments which are considered unreliable for our classification purposes.

The improved clone sequences furnished by the browser are a direct representation of GenBank accession records. As such, the coordinates of the sequences and their underlying features have not been altered in any way. The highly-repetitive nature of the sequences makes it difficult to elucidate further contig order and orientation at this time.

We have gone to great lengths to ensure consistency and to prevent any incongruence between the data available at NCBI and the data we provide. We hope to make our approach to analyzing the emerging maize genome sequence as transparent as possible.

As primary sequencing of the maize genome winds down and the sequence becomes more defined, we are concentrating our efforts to provide higher-quality annotation builds. We are integrating new and updated sources of evidence and refactoring protocols to add better context and meaning to the produced annotations.

MaizeSequence.org is a replacement the maize browser on Gramene. Because the maize browser and the annotation pipeline are an integral part of the Maize Genome Sequencing Project and track its progress, data such as clones and underlying annotations is produced more rapidly. The browser therefore warrants its own website. Nevertheless, all cereal features between Gramene and MaizeSequence.org are deeply linked, and users should be able to navigate seamlessly between different genome maps provided by both browsers. Aside from an extensive set of enhancements, users should feel no different.

We strive to provide as much information on this site as possible. If a question does not appear in this FAQ, please use our feedback form to contact us. We will try our best to respond. If the question is general enough, and it usually is, we will post it on the FAQ.

We freeze the sequences on MaizeSequence.org in order to run our annotation pipeline and to validate the results. Data checkpoints are taken to accommodate the evolving nature of the annotation pipeline, as more analyses are integrated that provide higher-order annotations. Data snapshots are taken on a quarterly basis. More information about the specific freezes and the release schedule are available on our new release page.

External clones are clones that have been sequenced through pilot projects that preceded the primary sequencing effort. There are about 450 such clones. Because the clones were sequenced under varying standards of quality, they are not being considered part of the minimal tiling path that defines the genome, and are not being re-sequenced as part of the project.