Figure 1.
Local assembly with RAD paired-end contig libraries.
(A) DNA fragments created by RAD tag library preparation have a restriction site (orange) and associated sequence (dark brown) at one end, and a random sheared-end sequence (light brown) at the other. (B) Paired-end sequencing of RAD tag libraries allows the assembly of the sheared-end sequences into contigs (C), one RAD site sequence at a time. The distance at which the random end sequence lies, and hence the length of the contigs assembled, is dictated by the size of fragments isolated during the gel extraction step in the protocol.
Figure 2.
SNP identification using RAD paired-end contigs and confirmation with Sanger sequencing in stickleback.
(A) An example pileup of the sheared-end reads at one RAD site from two stickleback individuals with different lateral plate phenotypes. The contig built off the reads from both individuals is shown at the bottom of each pileup, with the reads from each individual placed above the assembly. A zoom in on the region containing a SNP in one individual (bold box) is shown in (B) with the polymorphic nucleotide highlighted in the box. (C) Histogram of contig length. The N50 length is 407 nucleotides. (D) Validation of SNP calls by Sanger sequencing of the region surrounding the SNP identified in (B). The high plate individual was confirmed to be heterozygous at the nucleotide position identified by the analysis, while the low plate individual was confirmed homozygous and matches the consensus at that position. (E) Verified RAD haplotype alleles: an example bi-allelic RAD sequence identified in the low plate individual (shown to the left) and the relevant contig region (right). The SbfI site is underlined, the SNPs confirmed by Sanger sequencing of individual amplicon clones are in bold. Nucleotide positions of polymorphisms, relative to the start of the SbfI site in the reference sequence, are displayed below.
Figure 3.
Sequencing and local assembly of overlapping contigs from stickleback fosmids using partial-digest RAD paired-end libraries.
(A) Incomplete digestion of DNA with a frequently cutting restriction enzyme creates overlapping restriction fragments. Preparing RAD-PE libraries from a stickleback fosmid following partial digestion with two frequent cutters resulted in contigs up to 1000 bp long and an N50 individual contig length of 481 nucleotides. (B) Shows the distribution of contigs built from the two libraries. (C) Mapping the contigs (black bars) from each RAD site sequence (grey boxes) back to the stickleback reference sequence demonstrated overlapping coverage over an ∼40 kb stretch of the genome with a zoom on part of the assembly displayed below.
Figure 4.
Whole genome sequencing and local assembly of overlapping contigs with partial-digest RAD paired-end libraries in E. coli.
Partial-digest RAD-PE libraries created individually assembled contigs up to 1300 bp long with an N50 length of 649 nucleotides (A), or 729 nucleotides if only the longest contig was chosen from each RAD site assembly (B). (C) RAD site sequences (grey boxes) and associated contigs (black rectangles) shown for a 50 kb region of the reference genome (annotated gene regions shown as thicker black rectangles). A close-up of a smaller region is at bottom.
Figure 5.
Long-insert RAD paired-end contigs.
Increasing the length of RAD fragments isolated before paired-end library preparation (A) and adding two circularization steps (B, C) creates short fragments with two ends that were originally distantly separated. (D) An example of the pileup of reads (grey bars) and the resulting contigs (blue and green lines) from both sides of one SbfI site (short red bars) in E. coli assembled from a long-insert RAD library. The individually-assembled contigs produced were up to 5 kb in length with an N50 length of 3,807 base pairs (E).