Corrigendum: RNA partitioning into stress granules is based on the summation of multiple interactions

RNA 27: 174–189 (2021)

The authors wish to correct an error in one of our paired RNA-seq samples. Specifically, one of the reported pairs of RNA-seq data sets from ΔΔG3BP1/2 PABPC1 purified stress granules was misannotated as ΔΔG3BP1/2, when in fact it was a wild-type sample.

Inspection of IGV traces for the ΔΔG3BP1/2 replicates showed that SRR7830181 (a SG purified sample) and SRR7830184 (a total RNA sample) did not exhibit a pattern consistent with the expected knockout pattern (Kedersha et al. 2016). These observations indicate that SRR7830181 and SRR7830184 RNA-seq data sets were mislabeled as a ΔΔG3BP1/2 knockout, and were in fact WT.

We addressed whether this errant sample impacted the analysis in Figure 7B in our original article. We reanalyzed our data by both including and excluding the errant samples. Two observations indicate that the errant sample does not affect the major conclusions drawn in our original article. First, Pearson's correlation coefficients for the original figure are comparable when we include (ρ = 0.93) and exclude (ρ = 0.92) the errant sample (see Figs. A and B). Second, when we compare SG enrichment from the two previous analyses, we see high degrees of correlation (ρ = 0.98, see Fig. C).

(A) Scatterplot depicting sorbitol-induced PABPC1-purified SG enrichment in ΔΔG3BP1/2 versus WT lines. The misannotated ΔΔG3BP1/2 sample is included in this analysis. (B) Same as A but excluding the errant sample. (C) Scatterplot depicting SG enrichment in corrected analysis versus previous errant analysis.

Due to the high degree of similarity between the updated analysis and the original findings, we do not believe this errant sample impacted any of the original findings, but Figure 7B does need to be replaced by Figure B above and has been corrected online. Also, these samples have been removed from NCBI GEO.

doi: 10.1261/rna.079790.123