Technologies to quickly and cost-effectively identify single nucleotide polymorphisms (SNPs) or other genetic/epigenetic variations at large scale is critical in understanding the relationship between genotypes and phenotypes. Application of such technologies will find wide application ranging from medical genomics, metabolic engineering, to evolutionary study and synthetic biology. DNA sequencing is regarded as the "Gold Standard" to identify such sequence variations. The advent of a new generation of sequencing technologies has opened up many opportunities for the development of such applications. Conventionally, large scale targeted sequencing and whole genome sequencing have been done by Sanger sequencing of PCR amplicons, tiling arrays and mass spectrometry. However, in large-scale studies, PCR has a major limitation in scalability: it is difficult to perform a high degree of multiplexing in a single tube reactions. Tiling arrays and mass spectrometry require several steps including re-sequencing of candidate regions and cost more to confirm true positive and negative. In this study, we explore the capability of the Illumina Genome Analyzer (Solexa 1G) to deep sequence eleven laboratory evolved bacteria strains. Using only one out of eight channels at the cost of only a few hundred dollars, we achieved more than 20X coverage for each strain and identified previous known single nucleotide changes and in high confidence possible new ones, and we also identified possible deletion/insertion regions. To take full advantage of the power of next-generation DNA sequencing methods, we also developed a new cost effective method for highly multiplex amplification of arbitrary sets of short sequences from a complex genome. We demonstrated the feasibility of simultaneously capturing over 10,000 target exons single tube reactions with ~98% specificity. One current limitation of the current Genome Analyzer is that it only reads sequences from one end of the constructs. Developing a pair-end sequencing capability will dramatically expand our ability to detect insertion, deletion, copy number variation and genome rearrangement.