Four of the signals listed in Table 2 were testable in the ACS data, and details of the replication analysis are summarized in Table 3. The format is similar to that of Table 2. Our criteria for replication are that the RPM model accounts for at least a 15% increase in explained variance and that the RPM models produced in the two datasets should be consistent. Three of the four signals replicate. All of these signals displayed >20% increase in explained variance, well above our threshold for replication. For three of the four, the RPM models in the ACS data were similar to those found for the COGEND data, meeting our criteria for replication. The fourth RPM model was not sufficiently similar to the one proposed in the COGEND data for us to consider it to be a replication.
