Using linkage maps to correct and scaffold de novo genome assemblies: methods, challenges, and computational tools

Abstract

Modern high-throughput DNA sequencing has made it possible to inexpensively produce genome sequences, but in practice many of these draft genomes are fragmented and incomplete. Genetic linkage maps based on recombination rates between physical markers have been used in biology for over 100 years and a linkage map, when paired with a de novo sequencing project, can resolve mis-assemblies and anchor chromosome-scale sequences. Here, I summarize the methodology behind integrating de novo assemblies and genetic linkage maps, outline the current challenges, review the available software tools, and discuss new mapping technologies.

Keywords: draft genome; next-generation sequencing; optical mapping; physical mapping; scaffolds.

Figure 1

(A) In whole genome assembly errors result from residual alleles which appear as discrete sequences in the reference, and mis-joins. Small fragments have no genomic context and contribute little information. (B) Using a genetic linkage map to anchor a de novo assembly resolves error in the reference sequence by giving small sequences genomic context, resolving allelism, and identifying mis-joins. Chromosome-scale assemblies can be constructed by ordering and orienting sequences with the linkage map. (C) A genetic linkage map can be estimated from a parental cross resulting in an F2, F3, or Backcross (here, BC1) population. Estimating a genetic linkage map requires (D) genotyping individuals at discrete markers (here, six markers across eight individuals with missing data); and (E) grouping markers into linkage groups; and ordering and spacing markers within linkage groups. Estimating order and spacing is difficult due to missing data and little recombination between adjacent markers.