In the seventies, Pete Riegel published a useful algorithm that related any time/distance to the (then-current) WR, as a way to equalize performance over different distances. It was a handy thing, and a bunch of us used to use it. For me (at age 38, oh so long ago):
10k 35:20 index = 781
10M 59:00 index = 777
but my best marathon was 2:51, index = 750. I always thought I could reach into the mid 2:40s somewhere (but never did).
A 2:45 would have been 778.
My other performances from 3-4 miles to 20K were consistently 775-785.
So if we use this procedure on the OP's time:
2:55 marathon = 733
37:40 10K = 733
I'm sure I don't have to point out the perils in using a single index that can't take training and conditions into account. But for all that, it gives you a reasonable yardstick, I think.