After the publication of this work , we became aware that AFDIL data set used to construct our database of sub-Saharan mtDNA sequences had been mislabelled, and in fact, did not contain Sierra Leone mtDNA sequences. We have obtained the correct Sierra Leone data set from AFDIL, reconstructed the database using the new file, and reanalyzed all of the data. The size of our database was reduced from 3725 to 3717 since the new Sierra Leone data set contained 109 sequences instead of 117 in the mislabelled data set. The swapping of data sets resulted in a cascade of minor corrections to Tables 2, 3, 4, 5, and 6 (see below). We also corrected an error in the entry for BAM013 in Table 7. However, the only major change was that there were a number of matches of African-American sequences to sequences in the mislabelled data set. Thus, the number of African-American matches to single ethnic groups dropped from a total of 16 to 9 (Table 3) representing just 5% of the African American sequences that we compared to the database.
Table 2. Characteristics of the sub-Saharan mtDNA HVS-I database
Table 3. Number of Perfect Matches to African-American HVS-I sequences
Table 4. Imperfect matches to the Gullah/Geechee and AFDIL African-American HVS-I sequences
Table 5. Geographical source of mtDNA HVS-I matches
Table 6. Distribution of single region matches
Table 7. Mitochondrial DNA HVS-I sequences included in this study
Table 8. Malinke and Bambara Mitochondrial DNA HVS-I sequences included in this study