Massively Parallel Functional Analysis of BRCA1 RING Domain Variants

doi:10.1534/genetics.117.300355

In the article by L. Starita, D. Young, M. Islam, et al. (GENETICS 200: 412–422) entitled “Massively Parallel Functional Analysis of BRCA1 RING Domain Variants,” there was a discrepancy in the data presented in Supplemental Material, Table S2 and in Figure S4, in regard to the effect of mutations in amino acids 104–304 on E3 ligase activity. The cause of this discrepancy was the inclusion of data points in Figure S4 that did not pass the read count filter and should have been eliminated, as they were for Table S2. The corrected Figure S4 contains information that now matches that in the original Table S2.

Additionally, on p. 415 in the second full paragraph, there was an error in the sentence beginning “Using an input frequency threshold (Figure S3A),” which has been corrected to read as follows: “Using an input frequency threshold (Figure S3A), we filtered these to a high-confidence set corresponding to 3170 amino acid substitutions, with the six replicates having Spearman’s rank correlation values between 0.76 and 0.83 (Figure S3B).”

Because the two-hybrid analysis of the RING domain focused only on residues 2–103, and all subsequent analysis, model fitting, and comparison of computational predictions in the article using either the E3 ligase or two-hybrid assay were confined to residues 2–103 of the RING domain, the change in Figure S4 does not affect any of the conclusions of this work.