Is it possible that they ruled him out with a sample submitted to ancestry by a relative from his adopted fathers side? Missing that Richard Allen was adopted?
Adopted or having a stepfather? With a known father or not? OK, let us assume, adopted.
Allen being a common name, I know a few, not necessarily from Indiana. Are all Allens related? Not paternally. However, given where the ancestors came from (England, Scotland perhaps slightly, Ireland), and how many Allens are there at Midwest, the chance to find Allens who would be related in Gedmatch is very high.
But what happens if Rick Allen is adopted? And it is not an NPE parent who later adopted (as such things happen), but a truly unrelated man. So they run BG's (unknown man's) DNA through Gedmatch and there might be a hit with some 6th cousin Allen who randomly matches unknown DNA (because at the level of 6th cousins lots of people might be related), but not enough to work with. And if Indiana Allens have the same Y haplotype, say, R1B1, and RA has, imagine, I haplotype, then the answer would be, "he is not Allen".
Much depends on biodad. Supposedly, he is living in the US and his relatives are in the system. So genetic genealogists make trees, find some "potential matches", maybe even work down the tree, collect DNAs from someone living close at Midwest, but all these "possible" matches are misses. Because RA is not in those trees.
Or, even more complicated. Imagine, biodad lives abroad. Not only may he not be in the system. I Googled countries 23@me ships to. EU mostly, but for some odd reason, even France is off the list. Mexico is off the list. All Latin America is off. Turkey... you name it. (maybe it has to do with genetic laws in these countries). Any US expat living abroad might at least have relatives in the US. But a student on a sabbatical in the US who had a kid and then moved back to France or Argentina... chances are, his relatives won't be in Gedmatch. Gedmatch is mostly fed by people living in the US.
Now there is also the other line, the maternal one. Lots of things to find there, maybe more. All is possible if she is a local, US-born, if her relatives are cultured, interested in genetics, believe in DNA. If they have money for these tests. A lot depends on how common her mitogroup is.
If they don't have a very good DNA to start with, and part of this DNA is from someone not in Gedmatch, then, yes, it could be a hassle.
Another situation. What if LE had no nuclear DNA, but just a hair from the CS? Nowadays people learned how to extract nuclear DNA from hair, but I suspect, the quality of the material varies. And plus they didn’t do it in 2017 yet. But, mitochondrial DNA is abundant. How much can be gleaned from it? Something, if women in her maternal line did not die childless. But imagine this - RA’s mom has direct maternal relatives with kids, and hers mito DNA is common, say, H, or HV. But mitochondrial DNA mutates. So in the unhappy circumstance the mutation happened in RA's mom, her DNA is H with a rare private mutation, and no one in Indiana matches her! And she is not into genealogy. It happens.
There are so many options why it didn’t work.