A Comparison of Whole-Genome Shotgun-Derived Mouse Chromosome 16 and the Human Genome

Venter J. C., et al., Science 291, 1304 (2001).

Lander E. S., et al., Nature 409, 860 (2001).

Gurwitz D., Weizman A., Drug Discov. Today 6, 766 (2001).

Adams M. D., et al., Science 287, 2185 (2000).

Myers E. W., et al., Science 287, 2196 (2000).

Rubin G. M., Lewis E. B., Science 287, 2216 (2000).

Hattori M., et al., Nature 405, 311 (2000).

The strains used in this study originate in whole or in part from mice obtained from A. Lathrop, who maintained a mouse farm in Granby, MA (53b). Her mice served as progenitors for many of the inbred strains that are in use today. The founders of the A/J strain arose from a cross of the Cold Spring Harbor albino stock and the Bagg albino stock in 1921 by Strong (53b). This strain is homozygous for mutant alleles at the agouti [a; nonagouti (black) allele of the agouti gene a; chromosome 2); brown (b; the brown allele of the tyrosine-related protein gene Trp1; chromosome 4); and albino (c; the albino allele of the tyrosinase gene Tyr; chromosome 7)] genes. DNA was prepared from mice that were obtained from the Jackson Laboratory (Bar Harbor, Maine). A/J mice are highly susceptible to cortisone-induced cleft palate; they show a high incidence of spontaneous and carcinogen-induced lung adenomas in multiparous females; they have a defect in macrophage function; and they fail to develop atherosclerotic lesions when fed an atherogenic diet. A/J mice are highly susceptible to lung and mammary cancer and have a moderate susceptibility to lymphomas. DBA was the first strain to be inbred and was established from various coat color stocks by C. C. Little in 1909. In 1929–1930, several DBA sublines were crossed and new strains were created, two of which persist today as DBA/1 and DBA/2 (53b). This strain is homozygous for the mutant alleles at the agouti and brown genes, as well as the dilute allele of the myosin Va gene Myo5a (chromosome 9). DBA/2J is one of two DBA sublines maintained by the Jackson Laboratory. DNA was prepared from mice obtained from the Jackson Laboratory. DBA/2J mice are resistant to atherosclerotic lesions when fed an atherogenic diet; they show a high frequency of age-dependent hearing loss and glaucomes; they are susceptible to audiogenic seizures; and they are extremely intolerant of alcohol and morphine. DBA/2J mice are highly susceptible to mammary cancer and have a moderate susceptibility to intestinal, liver, and lung cancer. The history of the 129×1/SvJ and 129S1/SvImJ strains is complex and has been recently characterized (54–56). The 129 family of strains originated with L. C. Dunn in 1928, and they are derived from the 101 strain. The founders were a combination of coat-color stocks from English fanciers and a chinchilla c^ch stock from Castle (53b). 129S1/SvImJ was formerly called 129/Sv-p+Tyr+Kitl+/J (Jackson Laboratory repository number JR02448), and it is homozygous for the white belly agouti allele (Aw/Aw) of the agouti gene. 129×1/SvJ was formerly called 129/SvJ, and it is homozygous for the white belly agouti (Aw) mutation as well as for the pink-eyed dilution mutation; it is segregating for the albino (c) and the albino-chinchilla (c^ch) alleles of the tyrosinase gene. DNA was prepared from 129×1 mice and 129S1 mice that were obtained from The Jackson Laboratory. 129 mice are susceptible to spontaneous testicular teratocarcinomas, and they are moderately susceptible to liver, lung, and mammary cancer as well as lymphomas. C57BL/6J mice are susceptible to diet-induced obesity, type II diabetes, and atherosclerosis. They have low bone density, late-onset heading loss, and a high incidence of microphthalmia. They are resistant to audiogenic seizures and are prone to hydrocephalus, and they show a strong preference for alcohol and morphine. C57BL/6J mice are highly susceptible to pituitary cancer and lymphomas and moderately susceptible to liver cancer.

Lindbad-Toh K., et al., Genesis 31, 137 (2001).

H. Morse, in Origins of Inbred Strains, H. Morse, Ed. (Academic Press, New York, 1978), pp. 3–21.

H. Morse, in The Mouse in Biomedical Research, vol. 1, History, Genetics and Wild Mice, H. Foster, J. Fox, J. Small, Eds. (Academic Press, New York, 1981), pp. 1–16.

Plasmid DNA libraries of three size classes (2, 10, and 50 kbp) were constructed and sequenced as described (1). DNA was isolated from spleen (71%), liver (23%), and kidneys (6%) of the following mouse strains obtained from Jackson Laboratories: 129×1/SvJ, DBA/2A, A/J, 129S3/SvImJ. Only tissues from male mice were used. Tissues were homogenized in a solution containing 1.0% SDS, 1× SSC, and 0.25 mg of Pronase E (Sigma) per ml and incubated at 37°C for 30 min. The homogenized tissue was then extracted once with an equal volume of aqueous phenol and two times with chloroform-isoamyl alcohol (24:1). DNA was precipitated with ethanol and dissolved in 1 ml of TE buffer. Plasmid library construction and DNA sequencing were performed as described (1). A total of 27.4 million successful sequence reads were collected between May and December 2000, and the assembly presented here was completed in March 2001.

R. J. Mural et al., unpublished data. 13a.  Whitehead genetic and RH STS markers were searched by ePCR and BLAST. A modified version of ePCR was used, allowing one gap and/or one mismatch 5′ of the 7th bp from the 3′ end. The full sequence of those STSs whose primers were not found by ePCR was searched against scaffolds by BLAST, at a 90% identity and 90% length match cutoff. The best matching primer or sequence location was used for subsequent mapping. 13b. 

Van Etten W. J., et al., Nature Genet. 22, 384 (1999).

BAC end sequences (BES from The Institute for Genomic Research) were mapped onto scaffolds by BLAST. Consistently oriented BES pairs between scaffolds are used to link the scaffolds to form a chain of scaffolds. Internal scaffolds recruited by this method and without mapping STS markers are considered bounded.

Footz T. K., et al., Genome Res. 11, 1053 (2001).

Scaffolds containing >10 kbp of sequence were analyzed for features of biological importance through a series of computational steps, and the results were stored in a relational database. For scaffolds >1 megabase, the sequence was cut into single megabase pieces before computational analysis. All sequences were masked for complex repeats with Repeatmasker before gene finding or homology-based analysis. The computational pipeline required ∼7 h of central processing unit (CPU) time per megabase, including repeat masking, or a total compute time of about 20,000 CPU hours. Protein searches were performed against the nonredundant protein database available at the National Center for Biotechnology Information (NCBI) and translations of the Celera Human Transcript data set. Nucleotide searches were performed against mouse, rat, and human Celera Gene Indices (assemblies of cDNA and EST sequences); mouse, rat, and human EST data sets parsed from the dbEST database (NCBI); mouse EST sequences from the Riken Mouse Encyclopedia sequence data archive (www.genome.rtc.riken.go.jp/); rat and human mRNA data sets parsed from the RefSeq experimental mRNA database (NCBI); the Celera Human Transcript data set; dog genomic DNA reads generated at Celera (1.5×); and pufferfish genomic DNA reads in the public domain, including Tetraodon nigroviridis and Fugu rubripes (NCBI). Initial searches were performed on repeat-masked sequence with BLAST 2.0 (57) optimized for the Compaq Alpha compute-server and an effective search space size of 3 × 10⁹ for BLASTN searches and 1 × 10⁹ for BLASTX searches. Additional processing of each query-subject pair was performed to improve the alignments. All protein BLAST results having an expectation score of <1 × 10⁻⁴, mouse nucleotide BLAST results having an expectation score of <1 × 10⁻⁸ with >94% identity, and human and rat nucleotide BLAST results having an expectation score of <1 × 10⁻⁴ with >80% identity were then examined on the basis of their high-scoring pair (HSP) coordinates on the scaffold to remove redundant hits, retaining hits that supported possible alternative splicing. For BLASTX searches, analysis was performed separately for selected model organisms (yeast, mouse, human, Caenorhabditis elegans, and D. melanogaster) so as not to exclude HSPs from these organisms that support the same gene structure. Sequences producing BLAST hits judged to be informative, nonredundant, and sufficiently similar to the scaffold sequence were then realigned to the genomic sequence with Sim4 for ESTs and mRNAs, and with Genewise for proteins. Because both of these algorithms take splicing into account, the resulting alignments usually give a better representation of intron-exon boundaries than standard BLAST analyses and thus facilitate further annotation (both machine and human). In addition to the homology-based analysis described above, three ab initio gene prediction programs were used (58–61a).

Corresponds to genes with two to four lines of evidence, as described in Table 8 of (1).

Maglott D. R., Katz K. S., Sicotte H., Pruitt K. D., Nucleic Acids Res. 28, 126 (2000).

Makalowski W., Boguski M. S., Proc. Natl. Acad. Sci. U.S.A. 95, 9407 (1998).

Wasserman W. W., Palumbo M., Thompson W., Fickett J. W., Lawrence C. E., Nature Genet. 26, 225 (2000).

Frazer K. A., et al., Genome Res. 11, 1651 (2001).

Krivan W., Wasserman W. W., Genome Res. 11, 1559 (2001).

For the DNA level searches, an initial set of matching segments was obtained by comparing the human and mouse sequences using BLASTn, with standard settings and an expectation value cutoff of 1 × 10⁻¹⁰. The matches were further filtered by requiring that at least 70 bp of the alignment, or the length of the alignment if <70 bp, did not overlap in either the human or mouse sequence with any other match. Thus, all of the final anchors are apparently unique in both genomes. Moreover, most anchors probably reflect significant conservative pressures. Based on the frequency of synonymous substitutions, about 1 in 3 bases would show point substitutions in the absence of selection or insertion or deletion events (61b). Under a null hypothesis of point mutation only and fixation resulting from random drift, the expected interval between anchors can be estimated to be >50 kbp, based on a random walk Markov model with probability of a +1 step equal to two-thirds and a one-third chance of a −3 step, given that the minimum alignment score for any anchor was 38 (+1 and −3 are the default match and mismatch costs for BLASTn searches). In contrast, we observed anchors approximately every 9 kbp. Furthermore, given a minimum score of 38 and the same random walk model, <5% of matches would be expected to receive a score of at least 52, whereas 79% of observed anchors achieved this level. Some of the anchors may nonetheless be due to chance preservation of ancestral sequence. This does not mean that such an anchor is invalid (the human and mouse sequences are ultimately orthologous), but only that the similarity is not always a sign of conservative evolutionary pressures.

The sex chromosomes are underrepresented in a whole-genome shotgun assembly and hence are not appropriate substrates for this type of analysis.

A run of consistent anchors is considered to be inverted if, by inverting that run of anchors on one genome, it could be merged with another run to one side or the other. For instance, if a set of anchors occurs as ABCDEF in human and as ABD′C′EF in mouse, where D′ and C′ are in reverse orientation, then the CD/D′C′ is inverted. A run of consistent anchors is considered to be locally transposed if exchanging it with another run to one side in one genome would allow the two runs to be merged. For instance, if a set of anchors occurs as ABCD in human but as CDAB in mouse, then exchanging CD and AB in the mouse would result in all four anchors belonging to the same run.

Lund J., et al., Mamm. Genome 10, 438 (1999).

Lund J., et al., Genomics 63, 374 (2000).

Antona V., et al., Cytogenet. Cell Genet. 83, 90 (1998).

Lengeling A., et al., Genet. Res. 66, 175 (1995).

Reeves R. H., Rue E., Yu J., Kao F. T., Genomics 49, 156 (1998).

Deng Z., et al., Genomics 18, 156 (1993).

Delcher A. L., et al., Nucleic Acids Res. 27, 2369 (1999).

Three was picked empirically.

MUMmer was run with a minimum word match length of 11 amino acids. At least three distinct protein matches were required, within a 15 “protein call” span, in order to be recognized as a candidate syntenic block. The entire mouse proteome was compared with the human proteome by MUMmer in 3 h of computation. In addition, all predicted human proteins were compared with predicted mouse proteins with BLASTP. Proteins were considered orthologous only if they were among the top five BLASTP hits in the target genome and had a BLAST expectation value of <10⁻¹⁰. Proteins scored as orthologs were then clustered into syntenic blocks in the same manner as for the MUMmer matches. Results derived from the two comparisons were merged in order to include the maximal number of orthologous proteins in the syntenic blocks. By using more sensitive MUMmer parameters, we identified additional syntenic blocks, corresponding to ancient human intragenomic duplications reported in (1).

For the purposes of this paper, we assert genes or proteins to be orthologous only if they are mutual best matches by BLAST and map to regions of conserved synteny, as determined by DNA sequence–based anchors (see text), between the mouse and human genomes. In addition, we include in our analysis only the class of paralogs that are the result of local gene duplication in one or the other species. We are aware that this is a narrow view of these concepts but use it here for the sake of clarity.

Auffray J. C., Fontanillas P., Catalan J., Britton-Davidian J., J. Hered. 92, 23 (2001).

Reeves R. H., et al., Nature Genet. 11, 177 (1995).

Sago H., et al., Proc. Natl. Acad. Sci. U.S.A. 95, 6256 (1998).

Pletcher M. T., Wiltshire T., Cabin D. E., Villanueva M., Reeves R. H., Genomics 74, 45 (2001).

From the Jackson Laboratory maps, NCBI site maps, Oxford grids, and Annual Human Genome Project Report maps (www.informatics.jax.org/menus/map_menu.shtml, www.ncbi.nlm.nih.gov/Homology/, www.informatics.jax.org/searches/oxfordgrid_form.shtml, and www.ornl.gov/hgmis/research/mapping.html, respectively).

Doyle J. L., DeSilva U., Miller W., Green E. D., Cytogenet. Cell Genet. 90, 285 (2000).

Dehal P., et al., Science 293, 104 (2001).

Koop B. F., Trends Genet. 11, 367 (1995).

Zoubak S., Clay O., Bernardi G., Gene 174, 95 (1996).

S. J. O'Brien et al., Science 286, 458 (1999).

B. John, G. L. Miklos, Eukaryote Genome in Development and Evolution (Allen & Unwin, Boston, 1988).

Nadeau J. H., Taylor B. A., Proc. Natl. Acad. Sci. U.S.A. 81, 814 (1984).

Nadeau J. H., Sankoff D., Nature Genet. 15, 6 (1997).

Mallon A. M., et al., Genome Res. 10, 758 (2000).

Britten R. J., Davidson E. H., Q. Rev. Biol. 46, 111 (1971).

Soriano P., Macaya G., Bernardi G., Eur. J. Biochem. 115, 235 (1981).

Griffiths D. J., Genome Biol. 2, 1017 (2001);

. 53a.  A number of papers have been published by users of the Celera mouse data (62–65). 53b.  M. F. Festing, in Genetic Variants and Strains of the Laboratory Mouse, M. Lyon et al., Eds. (Oxford Univ. Press, Oxford, 1996), pp. 1537–1576.

Simpson E. M., et al., Nature Genet. 16, 19 (1997).

Threadgill D. W., et al., Mamm. Genome 8, 390 (1997).

Festing M. F., et al., Mamm. Genome 10, 836 (1999).

Altschul S. F., Gish W., Miller W., Myers E. W., Lipman D. J., J. Mol. Biol. 215, 403 (1990).

Uberbacher E. C., Xu Y., Mural R. J., Methods Enzymol. 266, 259 (1996).

Burge C., Karlin S., J. Mol. Biol. 268, 78 (1997).

Mural R. J., Methods Enzymol. 303, 77 (1999);

. 61a. 

Salamov A. A., Solovyev V. V., Genome Res. 10, 516 (2000);

. 61b. 

Kumar S., Subramanian S., Proc. Natl. Acad. Sci. U.S.A. 99, 803 (2002).

I. Rodriguez et al. Nature Neurosci.5, 134 (2002).

Sierra D. A., et al., Genomics 79, 177 (2002).

Zhang X., Firestein S., Nature Neurosci. 5, 124 (2002).

Young J. M., et al., Hum. Mol. Genet. 11, 535 (2002).

Abril J. F., Guigo R., Bioinformatics 16, 743 (2000).

We thank additional members of Celera's DNA sequencing, software, and IT groups for their valuable contributions to this work.