Statistical approach to test and compare two sets of data that is disimilar in size?

Hello,
I am performing molecular docking on different protein targets that are divided into two groups. The result comes as an value and higher value is, better performance.
So I have data of different size (21 values, 20 values, 16 values and 10 values) and I am trying to compare values:
21 to 10 and
20 to 16.
I don't know much about statistics so if anybody can help me with suggesting approach that I could take to achieve my goal.
Thanks