R regex extract substring followed by end of line or specific character (lazy match)

Question

I have a vector mystr the elements of which contain the unit of measure for a given parameter - this is indicated by the letters, symbols etc. following UOM=<