how do you define "minority"?

I'm curious how others in this group would define "minority" in ACS or PEP terms if asked to do so. 

Background: We're running into an issue with some of our data tools which previously relied on single-year-of-age (SYA) bridged-race data from NCHS. Since those special tabs have been discontinued, I've started using the SYA "by special request only" data from the Population Estimates Program. Unfortunately, instead of the four bridged races in the NCHS data, the PEP data has 11 -- five each in the "alone" and "alone or in combination" groups, plus "two or more races". In both data collections, Hispanic is coded separately.

Previously we defined the "minority" as the total population minus the white non-Hispanic population. However, we now have to choose between two options (and there may be others):

1. total minus (white alone and non-Hispanic), or

2. total minus (white alone or in combination with other races and non-Hispanic).

As you can probably guess, option 1 gives a larger minority count, sometimes by as much as 4-5%. Also, the "alone or in combination groups" can overlap, so I'm concerned that using option 2 would result in some minority people being counted in the non-minority group.

Are there other approaches to defining "minority" that you use? I realize it can be a slippery concept, so I'm interested how others have tackled it in their own work.