Allelic ratios and the mutational landscape reveal biologically significant heterozygous SNVs

Abstract

The issue of heterozygosity continues to be a challenge in the analysis of genome sequences. In this article, we describe the use of allele ratios to distinguish biologically significant single-nucleotide variants from background noise. An application of this approach is the identification of lethal mutations in Caenorhabditis elegans essential genes, which must be maintained by the presence of a wild-type allele on a balancer. The h448 allele of let-504 is rescued by the duplication balancer sDp2. We readily identified the extent of the duplication when the percentage of read support for the lesion was between 70 and 80%. Examination of the EMS-induced changes throughout the genome revealed that these mutations exist in contiguous blocks. During early embryonic division in self-fertilizing C. elegans, alkylated guanines pair with thymines. As a result, EMS-induced changes become fixed as either G→A or C→T changes along the length of the chromosome. Thus, examination of the distribution of EMS-induced changes revealed the mutational and recombinational history of the chromosome, even generations later. We identified the mutational change responsible for the h448 mutation and sequenced PCR products for an additional four alleles, correlating let-504 with the DNA-coding region for an ortholog of a NFκB-activating protein, NKAP. Our results confirm that whole-genome sequencing is an efficient and inexpensive way of identifying nucleotide alterations responsible for lethal phenotypes and can be applied on a large scale to identify the molecular basis of essential genes.

Figure 1

A map of lethal genes on chromosome I exposed by the deletion hDf7, which is in the sDp2 region. Six lethal genes fall within this region. Three of these genes (let-353, let-503, and let-504) fall in the region flanked by the left breakpoint of hDf7 and the left breakpoint of the cosmid C18E3.

Figure 2

Positions of G→A (red squares) and C→T (blue diamonds) homozygous SNVs. Any SNV with >90% read support is considered homozygous. The x-axis represents the length of the chromosome in mega base pairs. The y-axis indicates each SNV ID. The scale is the same for all the chromosomes so that the slope of the line corresponds to the density of SNVs. Chromosomes I and III are predominantly C→T changes whereas chromosomes II, V, and X are predominantly G→A changes. Chromosome IV has a stretch of G→A changes followed by a stretch of C→T changes. Chromosome I and X have a steeper slope, indicating a higher density of EMS mutations.

Figure 3

(A) Alkylated G’s as a result of EMS treatment affecting the DNA of haploid gametes are shown in red. (B) During the first round of replication, the alkylated G’s will be mis-paired with T’s. (C) In the next round of replication, the new T’s will pair with A’s. Either one of the chromosomes shown in C could enter the P1 cell and give rise to the germ line and the next generation of offspring. If the chromosome giving rise to the germ line comes from the “New Minus” strand, the mutations will appear as G→A changes when compared to the reference sequence. If the chromosome giving rise to the germ line comes from the “New Plus” strand, the mutations will appear as a C→T changes. Other segregant possibilities are not shown for simplicity. Note that all the mutational changes along the chromosome are of the same type, either G→A or C→T, and that the type will be passed on to the progeny in the next generation.

Figure 4

The number of EMS-induced changes per 1 Mbp. This plot was generated by combining the rates from KR772 (excluding chromosome I), RB5002, VC1923, and VC1924. Two prominent peaks are clearly observed: one at 1.5/Mbp and another at 3.5/Mbp. A smaller peak was also observed at 6/Mbp. The data for RB5002, VC1923, and VC1924 are from Flibotte et al. (2010).

Figure 5

Allelic ratio in KR772 for the whole chromosome I, whole chromosome III, part of chromosome I under sDp2, and part of chromosome I not under sDp2. Allelic ratio is presented as the percentage of reads that show SNV at a particular nucleotide position. In the sDp2 region, the peak at 70–80% represents mutations homozygous in the homologs with a wild-type allele in sDp2.

Figure 6

Chromosomal I region between 6 and 9 Mbp. The blue bars represent nonhomozygous SNVs, and the red bars represent homozygous SNVs. An SNV with an allelic ratio between 40 and 89% is considered as nonhomozygous. An SNV with allelic ratio ≥90% is considered as homozygous. A nonhomozygous mutation first occurred at 7.3 Mbp.

Figure 7

Location of let-504 alleles. The changes underneath the allele name indicate amino acid changes.