EnGarde

I don't want to reveal identifying information, apart from that the CS program was at an R1 institution. Virtually all good CS master's programs are self-funded, and many have deadlines later than PhD deadlines, so this should work if you can't get an invitation from a Ph.D. program Choosing a master's program that's conveniently located for you (in case of in-person classes/research opportunities) or even better, in one of the same institutions one of your aspirational Ph.D. programs, may be most strategic. But you need to act quickly (and perhaps in parallel), especially if you're asking for an exception in their application procedure (fortunately this year's early GRFP announcement helps you do this). As a backup, reach out to your own institution to see if there's a master's program locally that you can enroll in. Note that you must be able to deploy your GRFP award within the field of study you applied to; however the program you join does not need to have the same name, as long as your first-year GRFP-sponsored research remains in original field (hence his ability to join a computational MS program with an award in neuroscience).

I agree about reaching out to programs you applied to with the news about your award. If you receive a reply that it's too late for their Ph.D. admissions, and since the award cannot be deferred, ask them whether you could begin within their MS program and then roll over into their Ph.D. program. You can also look around for a relevant MS program (anywhere), explain your dilemma, and ask for a timely reply. This worked for an ugrad from my institution who was switching fields for grad school (from physics into computational neuroscience) and was over-ambitious in applying to only the 3 top grad programs in the country in his new field; he hung out in a master's program in computer science for a year or two and then successfully transferred his award to a Ph.D. program. Good luck.

Check during the wee hours after midnight.

The current Solicitation says 1500 slots. This appears to be a hedge against budget cuts, which had been planned but ended up being restored by Congress. Note that the NSF budget for 2020 provides for only 1600 new awards. It would be a good idea to write to your Reps (especially those on Appropriations and on Science, Space, and Technology Committees) and ask them to advocate for more fellowships, not fewer.

NSF is bound to follow the solicitation, and if your actual research isn't disease related/biomedical, then your research plan should be eligible. In previous years, rejected applicants were given the option to appeal, so it should be no different this year (despite the claim that "this decision is final" - after all, sometimes mistakes are made). While you cannot appeal the final outcome of whether you win an award or not, I think you should respectfully appeal this ineligibility judgment.