Genetic mapping of egg-killing HR to a single PEK locus
through BSA-seq
To identify genomic regions associated with Pieris egg-induced
HR, we performed a B ulk s egregant a nalysis coupled
with whole genome s equencing (BSA-seq) on the backcross
population BC1-3 (Fig. 2b). We generated two bulks (n = 10) with
either susceptible plants (S-Bulk) or with resistant plants (R-Bulk).
Genomic DNA of S-bulk, R-bulk, the S parent (DG1-S1), the R parent
(F1_#130) and the donor of HR (SF48-O1, “R donor”) was sequenced
using Illumina 150 bp paired-end reads yielding between 14 and 22 Gb
data for each sample (Supporting information: Table S2). As we estimated
a haploid genome size of our B. nigra material of
~550 Mb, our sequencing resulted in a read depth between
26x and 35x for each diploid genome of parents, thus a coverage ranging
between 13x and 17.5x of a haploid genome. Given the heterozygosity in
our plant material, we performed a k- mer based BSA-seq approach
(CoSSA) which was recently developed for a highly heterozygous
tetraploid potato (Prodhomme et al. 2019). Our genetic model
pointed at a monogenic dominant locus which was heterozygous in the R
parent F1_#130, homozygous in the S parent DG1-S1 and segregating 1:1
in the backcross population BC1-3. Thus we expected a single resistant
(R) allele in the R parent conferring HR. Accordingly, we first
generated k -mer tables (k = 31) for each sample
independently. Then we selected k -mers from the R allele (“R
haplotype”) by using basic set algebra to retain k -mers that
were unique to the R-bulk and originated from R parent and R donor
(Supporting information: Fig. S3). The resulting R haplotype-specifick -mer set was filtered to retain unique k -mers with a
frequency similar to half of sequencing depth for a haploid genome (15x+ 5) to discard k -mers derived from repeated regions
(Supporting information: Table S3).
The unique R haplotype-specific k -mers were aligned to 1 Mb bins
of the B. nigra reference genome C2 (Supporting information: Fig.
S4). Approximately ~85% of the R-specific k -mers
were successfully aligned, while the rest likely represented sequences
from plant material that were too divergent or absent from the B.
nigra reference genome. A unique single peak consisting of
~73% of the R haplotype-specific k -mers was
found in a 10 Mb interval on the proximal end of chromosome B3, spanning
from 3 Mb to 13 Mb (Fig. 2c). All other k -mers
(~27%) mapped at a very low frequency (below 0.4% for
each 1 Mb bin) throughout the rest of the genome (Supporting
information: Fig. S4). Similarly, alignment of R haplotype-specifick -mers to other B. nigra genomes resulted also on a single
peak, namely on chromosome B3 of accession NI100 and chromosome B7 of
accession Sangam (not shown). Location of PEK on a different
chromosome in Sangam is likely due to a different chromosome naming by
the authors of the genome, as we found that chromosome B7 of Sangam is
perfectly syntenic to chromosome B3 of C2 and NI100 accessions
(Supporting information: Fig. S5). In conclusion, a BSA-seq approach
allowed us to map the HR cell death trait in a BC1population and identified a single genetic locus on chromosome B3. This
confirms the genetic model that we hypothesized based on the phenotypic
segregation.