Large-Scale Discovery of Gene-Enriched SNPs

Whole-genome association studies of complex traits in higher eukaryotes require a high density of single nucleotide polymorphism (SNP) markers at genome-wide coverage. To design high-throughput, multiplexed SNP genotyping assays, researchers must fi rst discover large numbers of SNPs by extensively resequencing multiple individuals or lines. For SNP discovery approaches using short read-lengths that nextgeneration DNA sequencing technologies offer, the highly repetitive and duplicated nature of large plant genomes presents additional challenges. Here, we describe a genomic library construction procedure that facilitates pyrosequencing of genic and low-copy regions in plant genomes, and a customized computational pipeline to analyze and assemble short reads (100–200 bp), identify allelic reference sequence comparisons, and call SNPs with a high degree of accuracy. With maize (Zea mays L.) as the test organism in a pilot experiment, the implementation of these methods resulted in the identifi cation of 126,683 putative SNPs between two maize inbred lines at an estimated false discovery rate (FDR) of 15.1%. We estimated rates of false SNP discovery using an internal control, and we validated these FDR rates with an external SNP dataset that was generated using locus-specifi c PCR amplifi cation and Sanger sequencing. These results show that this approach has wide applicability for effi ciently and accurately detecting geneenriched SNPs in large, complex plant genomes. THE AVERAGE NUCLEOTIDE diversity of coding regions between any two maize (Zea mays L.) lines (π = 1–1.4%) is twoto fi vefold higher than other domesticated grass crops (Buckler et al., 2001; Tenaillon et al., 2001; Wright et al., 2005). Moreover, it is not uncommon to fi nd maize haplotypes more than 2% diverged from one another (Tenaillon et al., 2001; Wright et al., 2005) and even as high as 5% (Henry and Damerval, 1997). Intragenic linkage disequilibrium (LD) rates rapidly decline to nominal levels within 2 kb in a population of diverse maize inbred lines (Remington et al., 2001). Of the ~2500 Mb that constitutes the maize genome, less than 25% is genic or low-copy-number sequence, with large blocks of highly repetitive DNA such as retrotransposons Published in The Plant Genome 2:121–133. Published 10 July 2009. doi: 10.3835/plantgenome2009.01.0002 © Crop Science Society of America 677 S. Segoe Rd., Madison, WI 53711 USA An open-access publication All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher. M.A. Gore, Dep. of Plant Breeding and Genetics, Cornell Univ., 175 Biotechnology Bldg., Ithaca, NY 14853; M.H. Wright, Dep. of Genetics and Development, Cornell Univ., 102 Weill Hall, Ithaca, NY 14853; E.S. Ersoz, Institute for Genomic Diversity, Cornell Univ., 175 Biotechnology Bldg., Ithaca, NY 14853; P. Bouffard, E.S. Szekeres, and T.P. Jarvie, 454 Life Sciences, 20 Commercial St., Branford, CT 06405; B.L. Hurwitz and A. Narechania, Cold Spring Harbor Lab., 1 Bungtown Rd., Cold Spring Harbor, NY 11724; T.T. Harkins, Roche Applied Science Corp., 9115 Hague Rd., Indianapolis, IN 46250; G.S. Grills, Life Sciences Core Labs. Center, Cornell Univ., 139 Biotechnology Bldg., Ithaca, NY 14853; D.H. Ware, USDA-ARS, Cold Spring Harbor Lab., 1 Bungtown Rd., Cold Spring Harbor, NY 11724; E.S. Buckler, USDA-ARS, Dep. of Plant Breeding and Genetics, Institute for Genomic Diversity, Cornell Univ., 159 Biotechnology Bldg., Ithaca, NY 14853. All custom code and scripts used in this study are available upon request from M. H. Wright ([email protected]). M.A. Gore and M.H. Wright contributed equally to this work. Received 14 Jan. 2009. *Corresponding authors ([email protected]) and