Corrigendum doi:10.1093/gbe/evu246 Selection on Horizontally Transferred and Duplicated Genes in Sinorhizobium (Ensifer), the Root-Nodule

The sentence in the Results at the top of page 1204 that begins, “For all three replicons in S. meliloti, DTdup was more negative than DT nonsyn... ” should read “For all three replicons combined, and for the chromosome and pSymA separately in S. meliloti, DTdup was more negative than DT nonsyn...” In addition, tables 1, 3, and supplementary table S3 contained some erroneous values. In table 1, the values for p and DT for S. meliloti SNPs were calculated based on 24 S. meliloti strains, rather than the 20 strains used for the rest of the results. Similarly, the S. meliloti row in table 3 presents results for 24 strains, rather than for the 20 strain subset. The numbers presented in table S3 were based on an earlier version of the duplication detection pipeline—we neglected to update this table when we updated our pipeline. Corrected values are presented below. As a consequence of these corrections, the sentence at the bottom of page 1204 that reads, “Based on the minor allele frequency spectrum, 10 % of nonsynonymous mutations in S. meliloti are effectively neutral, whereas two-thirds are subject to very strong (probably negative) selection (table 3) ” should read “Based on the minor allele frequency spectrum, 10 % of nonsynonymous mutations in S. meliloti are effectively neutral, whereas three-quarters are subject to very strong (probably negative) selection (table 3).” The corrected values are similar to the erroneous values and do not affect any of the other tables, figures, or results, and they do not alter our principal finding—that duplications and horizontally transferred genes tend to segregate at low frequency relative t