A shotgun optical map of the entire Plasmodium falciparum genome

1W.M. Keck Laboratory for Biomolecular Imaging, Department of Chemistry and 2Courant Institute of Mathematical Sciences, Department of Computer Science, New York University, Department of Chemistry, New York, New York, USA. 3Malaria Program, Naval Medical Research Center and 4The Institute for Genomic Research, Rockville, Maryland, USA. 5Present address: University of Wisconsin-Madison, Departments of Chemistry and Genetics, UW Biotechnology Center, Madison, Wisconsin, USA. Correspondence should be addressed to D.C.S. (e-mail: schwad01@mcrcr.med.nyu.edu). The unicellular parasite Plasmodium falciparum is the cause of human malaria, resulting in 1.7–2.5 million deaths each year1. To develop new means to treat or prevent malaria, the Malaria Genome Consortium was formed to sequence and annotate the entire 24.6-Mb genome2. The plan, already underway, is to sequence libraries created from chromosomal DNA separated by pulsed-field gel electrophoresis (PFGE). The AT-rich genome of P. falciparum presents problems in terms of reliable library construction and the relative paucity of dense physical markers or extensive genetic resources. To deal with these problems, we reasoned that a high-resolution, ordered restriction map covering the entire genome could serve as a scaffold for the alignment and verification of sequence contigs developed by members of the consortium. Thus optical mapping was advanced to use simply extracted, unfractionated genomic DNA as its principal substrate. Ordered restriction maps (BamHI and NheI) derived from single molecules were assembled into 14 deep contigs corresponding to the molecular karyotype determined by PFGE (ref. 3). Optical mapping is now a proven means for the construction of accurate, ordered restriction maps from ensembles of individual DNA molecules derived from a variety of clone types, including bacterial artificial chromosomes4 (BACs), yeast artificial chromosomes5 (YACs) and small insert clones6. We previously developed approaches for mapping clone DNA samples that relied on the analysis of large numbers of identical DNA molecules. Here, the challenge was to develop ways to generate restriction maps of a population of randomly sheared DNA molecules directly extracted from cells that were obviously non-identical. Problems to be solved included the development of techniques for mounting very large DNA molecules onto surfaces and new methods for accurately mapping individual molecules, which were uniquely represented within a population. Finally, new algorithms were necessary to assemble such maps into gap-free contigs covering all 14 chromosomes of the P. falciparum genome. We developed an optical mapping approach, termed shotgun optical mapping, that used large (250–3,000 kb), randomly sheared genomic DNA molecules as the substrate for map construction (Fig. 1a–e). Random fragmentation of genomic DNA occurred naturally as a consequence of careful pipetting and other manipulations. Surface-mounted molecules were digested using BamHI and NheI (refs 6–8). Because genomic DNA molecules frequently extended through multiple digital image fields, we developed an automated image acquisition system (GenCol) to overlap digital images with proper registration (Figs 1c and 2). Map construction techniques were altered to take into account local restriction endonuclease efficiencies (the rate of partial Fig. 1 Schematic of shotgun optical mapping approach. a, Shotgun optical mapping used large (250–3,000 kb), randomly sheared genomic DNA molecules as the substrate for map construction. b, Random fragmentation of genomic DNA occurred naturally as a consequence of careful pipetting and other manipulations. Surface-mounted molecules were digested using BamHI and NheI (ref. 8). c, Because genomic DNA molecules frequently extended through multiple image fields, an automated image acquisition system was developed (GenCol) and used to overlap images with proper registration. d, Map construction techniques take into account local restriction endonuclease efficiencies (the rate of partial digestion) and the analysis of molecule populations that differed in composition and mass. e, These steps were necessary to enable accurate construction of map contigs. a