A single-cell transcriptome atlas reveals paradoxical monoallelic expression of RNA editing and heterozygous SNPs

Both genomic mutations and RNA editing contribute to functional complexity and drive adaptive evolution. Single-cell profiling offers deep insight into the cis-regulatory mechanisms underlying these variations. Using 13 025 single-cell Smart-Seq libraries from whole-body Drosophila melanogaster, we unexpectedly found that 94.0% of adenosine-to-inosine RNA editing sites and 92.8% of heterozygous single nucleotide polymorphismss (SNPs) with sufficient "unique fragment support" exhibit binary expression (0 or 1) in a single cell. The genotypes of representative heterozygous SNPs were validated by Sanger sequencing. Meanwhile, binary RNA editing itself is logically questionable due to elusive mechanism, compromised condition specificity, and untenable heterozygote advantage. This fact that for most cases in Smart-Seq, only a single allele (out of the various haplotypes) is finally maintained per cell, raises the following concern. Regardless of the biological or technical explanations like monoallelic transcriptional burst, dropout, or amplification bias that might account for this binary expression pattern, our findings conservatively indicate that Smart-Seq may not be good at analyzing molecular diversity and that the results need to be interpreted with caution.