Page 143 - PC2019 Program & Proceedings

PLANT CANADA 2019

S71. RNA-Seq estimated gene abundance differences between Zea mays genotypes are strongly
affected by read mapping bias
Zhan, S.; J. Tosh; C. Griswold; L. Lukens
University of Guelph

RNA sequencing has been used widely to quantify genetic variation of gene expression. A potential
problem with this approach is that RNA-Seq reads carrying a reference allele of a polymorphic locus may
map correctly to a reference genome, but RNA-Seq reads carrying a non-reference allele may not or
incorrectly map. To test the effect of preferential alignment on transcript abundance estimates, we aligned
RNA-Seq reads from 105 individuals derived from a biparental cross between inbred lines B73 and
Mo17, to a B73 reference genome. There is strong evidence for transcript abundances estimates biased to
B73. B73 alleles at 83% of the genes with transcript levels that differ between B73 and Mo17 alleles lead
to higher gene expression estimates than do Mo17 alleles. The number of single nucleotide
polymorphisms between alleles strongly correlates with the magnitude of the B73 allele’s preferential
expression. Reducing the stringency of RNA-Seq read/reference alignment criteria has little effect on the
number of genes with higher B73 transcript abundances but reduces the magnitude of the B73- Mo17
difference for 80-85% of all genes. Finally, the frequency of B73 read alignment to genomic DNA
encoding mRNA untranslated regions is notably higher than that of Mo17. Our results reveal that reduced
read mapping due in part to SNPs and to differences in UTR regions contributes to strong differences in
genetic estimates of gene expression.

Shuhua Zhan (szhan@uoguelph.ca)

S72. A Tale of Two Genomes: Methylome and transcriptome profiling of Brassica napus seed
development
*
Khan, D. ; D. Ziegler; M. Belmonte
University of Manitoba

Canola (Brassica napus) is one of Canada’s most economically valuable crops. B. napus is an
allotetraploid comprised of two progenitor genomes (B. oleracea and B. rapa). We can take advantage of
methylome and transcriptome sequencing to better understand how gene regulatory networks contribute
to valuable seed traits such as vigor, size, and nutrient content. We profiled the methylome and
transcriptome of canola seeds during morphogenesis and seed maturation. Significant subgenome bias is
observed in leaves and seeds. Genes, promoters, and repetitive elements are all more heavily methylated
in the B. oleracea subgenome than the B. rapa subgenome. Unique suites of promoters are differentially
methylated in early and late seed development, targeting genes involved in development and carbon
metabolism. Transcription factors are hypomethylated to other protein coding genes, possibly permitting
fast transcriptional responses during seed development. We uncovered distinct transcriptional networks in
the B. oleracea and B. rapa subgenomes of canola, indicative of unique regulatory mechanisms between
the two subgenomes. Several bZIP transcription factors are identified as potential regulators of energy
metabolism and development in canola seeds. Knockdown of BnBZIP11 using RNA interference resulted
in a seed-lethal phenotype and knockdown of target genes predicted to be activated by BZIP11. Our
analysis provides new insight into the genomic architecture and mechanisms of regulation that underlie
the complex processes of seed development.

Deirdre Khan (deirdre.khan@umanitoba.ca)

Page 141 of 339