I-TASSER server with automatic restraints? Result differs from the version in my computer

Hello.

I've been running I-TASSER using a very short sequence. I have obtained the results from the Zhang Lab's server and from my computer.

When I was browsing the results from the Zhang Lab's server through the Web Interface, I noticed that it had the section that looked as the following: "Top 10 templates used by I-TASSER". Under this section, it describes ten templates as follows:

1 1ba6A 1.00 1.00 1.00 3.75 Download DAEFRHDSGYEVHHQKLVFFAEDVGSNK
2 1bjbA 0.96 0.96 1.00 1.06 Download DAEFRHDSGYEVHHQELVFFAEDVGSNK
3 1ba6A 1.00 1.00 1.00 2.41 Download DAEFRHDSGYEVHHQKLVFFAEDVGSNK
4 1ambA 1.00 1.00 1.00 2.12 Download DAEFRHDSGYEVHHQKLVFFAEDVGSNK
5 1ambA 1.00 1.00 1.00 1.60 Download DAEFRHDSGYEVHHQKLVFFAEDVGSNK
6 1amlA 1.00 1.00 1.00 5.23 Download DAEFRHDSGYEVHHQKLVFFAEDVGSNK
7 1ambA 1.00 1.00 0.96 2.81 Download DAEFRHDSGYEVHHQKLVFFAEDVGSN-
8 1amlA 1.00 1.00 1.00 0.88 Download DAEFRHDSGYEVHHQKLVFFAEDVGSNK
9 1amlA 1.00 1.00 1.00 3.37 Download DAEFRHDSGYEVHHQKLVFFAEDVGSNK
10 2lp1A

However, I don't know where do this templates come from since I have not declared them at all. Is it that the I-TASSER online adds templates automatically? Are these 10 templates above passed to the I-TASSER runtime as -restraint1 options?

My other question is related to the differences between the result I get from the I-TASSER running in the Zhang Lab and the one I have in my computer system. The difference is substantial. The model1.pdb coming from the Zhang Lab's server has almost twice the amount of atoms than the one I obtained with the I-TASSER I downloaded. I was possible into possible causes of difference before posting this question, and it seems that the libraries really have nothing to do with it. Can somebody please explain me why results are so divergent?

Let me know.

Good day.