Leveraging Long-Read Low-Pass Sequencing for High-Resolution Trait Mapping in Peanut Breeding

Accelerating crop improvement is critical to meeting food security demands in a changing climate. Long-read sequencing offers advantages over short reads in resolving structural variations (SVs) and aligning complex genomes, but its high cost has limited adoption in breeding programs. This study explores long-read low-pass (LRLP) sequencing as a cost-effective alternative for genomic analysis in a peanut multi-parent advanced generation intercross. Here we analyze LRLP using a variety of methodologies including Khufu on a linear graph, pangraph, and dynamic pangraph, as well as open source tools to analyze SVs and coverage. Consistently we find increased variants called for LRLP data compared to short read data. With a 1.63x average depth, LRLP sequencing covered 55% of the genome and 58% of gene space, outperforming short-read sequencing, which achieved only 17% and 11%, respectively even at a depth of 1.68x. Enhanced alignment accuracy and data retention further demonstrated LRLP’s efficacy.Our results highlight LRLP sequencing as a scalable, cost-effective tool for high-resolution trait mapping, with transformative potential for plant breeding and broader genomic applications.************* Files******************Short_linear_09.hapmap- A haplotype map of 125 individual peanut samples derived from a MAGIC breeding population. The file represents SNPs and was produced by Khufu, a proprietary genotyping pipeline. The input data for this analysis was low-depth short-read sequences.Long_linear_09.hapmap-A haplotype map of 125 individual peanut samples derived from a MAGIC breeding population. The file represents SNPs and was produced by Khufu, a proprietary genotyping pipeline. The input data for this analysis was low-depth long-read sequences.Short_linear_09_imputation_eval.txt- A file produced by Khufu that scores the accuracy of imputation.Long_linear_impuation_eval.txt- A file produced by Khufu that scores the accuracy of imputation.Long.panmap- A variant map of 125 individual peanut samples derived from a MAGIC breeding population produced by KhufuPan, a proprietary genotyping pipeline using a pangenome graph as a reference. SNPs, indels, and SVs are represented in this panmap. The input data for this analysis was low-depth long-read sequences.Long.panmap.fa- A file produced by KhufuPan, a proprietary genotyping pipeline using a pangenome graph as a reference. This file corresponds to Long.panmap and shows the sequences of the variants. The input data for this analysis was low-depth long-read sequences.Short.panmap- A variant map of 125 individual peanut samples derived from a MAGIC breeding population produced by KhufuPan, a proprietary genotyping pipeline using a pangenome graph as a reference. SNPs, indels, and SVs are represented in this panmap. The input data for this analysis was low-depth short-read sequences.Long.panmap.fa- A file produced by KhufuPan, a proprietary genotyping pipeline using a pangenome graph as a reference. This file corresponds to Short.panmap and shows the sequences of the variants. The input data for this analysis was low-depth short-read sequences.LRLP_Sequencing_Statistics.xlsx- Raw sequencing statistics from PacBio's Revio Sequencer of all peanut samples.LRLP_PBSV_info.xlsx- Data on structural variants called in all LRLP data using PacBio's tool PBMM2 and PBSV.LRLP_snpeff_summary.html- link to snpEff results for LRLP peanut samples.CostCalculationInfo.xlsx- Cost breakdowns used to calculate cost per value.