The main concern to my mind is the pigeon blood gene, and while I like your proposal to create a strain of red eyed pigeons, a little bit of parental selecting is in order to make matters easier. Because each parent carried a single copy of the pigeon gene, most of the fry (75%) inherited this gene, with around a third of them inheriting both copies which makes them homozygous for that gene. These are the ones you want, because they will produce 100% pigeon fry, and when 2 homozygotes are bred together, then they not only produce 100% pigeons, but they will be 100% homozygous pigeons, a situation that is highly desirable. After this point, it is an easier matter of selecting for the traits you want, without having to worry about the integrity of the pigeon gene.
The tough bit is finding out which ones are the homozygotes you already have. You can't tell them apart just by looking, and crossing within the group is the most difficult way to do it as it won't lay the genes open for us to see. What is easiest is by crossing each potential parent with an unmutated wild type discus, something like a turk or a brown will do nicely. Because we know that the brown based discus does not carry a copy of the pigeon gene, when we cross we lay open the pigeon gene in the resulting fry. A homozygote pigeon parent, and every fry is a pigeon. A heterozygote pigeon parent, only some are pigeons.
You can do this with all the potential breeders including the red white ones to see if they carry both the red white and pigeon genes. There are so many potential directions to head in this sort of crossing, its a good idea to try and map your progress so you can better understand the results you will see.
Good luck and have fun, they look awesome!