Why are there differences in the species used for conservation between genome builds 37 and 38? This makes the significance of the region look very different. Is there not a specific combination of species that you use as standard? Do these alignments feed into the predictions that are automatically computed?

The Orthologue Alignments for each gene are downloaded from Ensembl Compara (https://www.ensembl.org/info/genome/compara/index.html). Differences between GRCh37 and GRCh38 are due to the species used in the alignments not being the same in all cases. The data for GRCh38 are more up to date in Ensembl, but it is extremely difficult to determine which alignment is better. AlamutTM Visual Plus includes a standard set of species, but this is dependent on what is available in Ensembl Compara. These sequence alignments are used for missense predictions in AlamutTM Visual Plus.