Accuracy of allele frequency estimation using pooled RNA-Seq

Area of Science:

• Population genetics
• Genomics
• Transcriptomics

Background:

• RNA-Sequencing (RNA-Seq) enables genome-wide variation analysis in nonmodel organisms via de novo transcriptome assembly.
• High costs of library preparation limit population genetic studies in nonmodel species.
• Sample pooling presents a cost-effective strategy for population-level analyses.

Purpose of the Study:

• To evaluate the accuracy of pooled RNA-Seq in estimating true allele frequencies.
• To compare pooled RNA-Seq accuracy with individually barcoded RNA-Seq.
• To identify factors influencing the accuracy of pooled RNA-Seq allele frequency estimation.

Main Methods:

• Analyzed liver transcriptomes of 10 bank voles, sequencing each sample individually and as a pool.
• Mapped RNA-Seq reads to a de novo assembled reference transcriptome.
• Compared allele frequencies from pooled samples to high-quality genotypes from individual samples (23,682 SNPs).

Main Results:

• Pooled RNA-Seq allele frequency estimates showed high correlation with true frequencies.
• Mean relative estimation error was 21%, independent of expression level.
• Minor allele frequency and interindividual gene expression variation impacted estimation accuracy.

Conclusions:

• Pooled RNA-Seq provides accurate allele frequency estimation, comparable to pooled genome resequencing.
• Assessing and accounting for interindividual gene expression variation is crucial for optimizing accuracy.
• This method is valuable for population genetic analyses of nonmodel organisms with cost constraints.