Table 4. Testing for the presence of significant conversion fragments within HLA genes (ordered according to their location on chromosome 6, from telomere to centromere) with the GENECONV program package.
| Locus | Number of sequences | Number of nucleotide sites1 | Number of polymorphic sites2 | Number of significant conversion fragments (no correction) | Number of significant conversion fragments (BLAST correction)3 | Number of significant conversion fragments (Bonferroni correction)4 |
| HLA-A | 414 | 546 | 157 | 23444 | 5 | 0 |
| HLA-Cw | 231 | 546 | 112 | 2145 | 2 | 0 |
| HLA-B | 755 | 546 | 160 | 102890 | 611 | 0 |
| HLA-DRB1 | 438 | 270 | 90 | 19917 | 15 | 0 |
| HLA-DQA1 | 33 | 249 | 45 | 179 | 1 | 0 |
| HLA-DQB1 | 70 | 270 | 37 | 164 | 0 | 0 |
| HLA-DPB1 | 121 | 264 | 43 | 2252 | 99 | 0 |
Computations are done with 0% of missing data allowed at each nucleotide site, and using all sequences described in the IMGT official database (release 2.13).
1
Exons 2 and 3 for class I genes, exon 2 for class II genes.
2
This represents the number of nucleotide sites used by the program for detecting putative conversion fragments.
3,4
Number of fragments remaining significant after corrections for multiple tests. The BLAST correction is described as more powerful than Bonferroni's correction by the author of GENECONV.