"Maize Assembled Genomic Island"


"Sorghum Assembled genoMic Island"


Like a woman's name (Maggie), NOT like a biblical wise man.


MAGIs are one of the publically available partial maize genome assemblies generated from Zea mays B73 genomic survey sequences (GSSs). Most of these GSSs were generated by the Consortium for Maize Genomics (Danforth Center, TIGR, Purdue University, and Orion Genomics) as part of an NSF project, and full details are available on the home page of this site.

The 114,173 MAGIs (version 3.1) have been subjected to both computational and biological quality assessments. Comparisons to maize BAC sequences suggest that over 95% (157/165) of MAGIs are correctly assembled and that the sequencing error rate in these contigs is <7.8 x 10^-4 (213/274,689 bp). Because the rates at which GSS junction-testing PCR primer pairs (~91%, 611/670) amplify genomic DNA are not significantly different than those of control primer pairs (~91%, 1,233/1,358), it appears that a very high percentage of genic MAGIs accurately reflect the structure of the maize genome (Fu et al., 2005). One cautionary note: we have demonstrated that about 40% of the high-Cot GSS clones included in the MAGIs contain transition mutations that appear to be cloning artifacts (Fu et al., 2004).

Assembly and data processing details are available in the Publications section of this site, as well as within the Methods summary.


SAMIs (Version 2) are contigs produced from ~50,000 methyl filtered (MF)sequence reads generated by by Drs. Martienssen and McCombie's groups at Cold Spring Harbor Laboratory and another 500,000 MF reads generated by Orion Genomics and their partners NC+Hybrids and Solvigen using Sorghum bicolor (BTx623) as source genomic DNA . These GSS sequences were assembled by the MAGI project using a non-uniform genome assembly strategy similar to that used for previous MAGI builds (Emrich et al., 2004; Fu et al., 2005).

Based on alignments to the sequences of BTx623 BACs (generated by the Bennetzen, Klein, Messing and Mullet labs) >96% (206/214) of the version 2.0 SAMIs appear to be correctly assembled and the sequencing error rate in these contigs is <2.4 x 10^-4 (56/230, 394 bp). The untrimmed vector or poor quality sequences at GSS termini were removed using the method of Fu et al., 2004.