Timing of repeated conversion events in the human α-globin cluster. Three sequential events occurred between paralogs containing the α3 pseudogene and the α2 gene. (A) The regions involved in each conversion are marked here with orange ovals on the self-alignment dot-plot for the human cluster, and labeled with their corresponding event id numbers. (B) The table displayed by Gmaj shows a summary of these events. The Event column (highlighted here with an orange oval) shows the assigned event id, and the Branch column (blue oval) indicates the estimated conversion time as a sub-lineage of numbered tree edges. The tree and colored numbers added here illustrate how to interpret the table and what information was used to infer the conversion time. Numbers in green are the ids of events found in the particular species (i.e. as primary), while those in brown indicate events found using that species as outgroup. Primary species are used to estimate the lower bound for the conversion time; e.g. we inferred that event 4 occurred before the split of human and chimpanzee because we observed this event in both species (in paralogous pairs that are orthologous to each other). In contrast, an outgroup species used for detecting an event provides evidence for the upper bound of its conversion time, i.e. the event occurred after the split of the primary and outgroup species. For example, event 5 must have happened after the separation of human and chimpanzee because it was detected when using chimpanzee as the outgroup and human as the primary species. CHAP does not draw any inference from negative results (lack of detection), since that may be due to other factors such as missing sequence.