I have performed a dynamics simulation of the same construct using two approaches.
1. A structure generated in the standard manner by reading the sequence from the psf file and then reading the coordinate information from the crd.
2 A structure generated by reading the sequence from a txt file of the sequence, Generating the sequence and then reading the crd file.

Is there an advantage in using either one or the other approach. The resulting simulations appear to be the same.

Thanks for any advice ahead of time.

At some point, someone used [2] to obtain the PSF used in [1]

For more complicated systems, reading a stored PSF is often easier, more efficient, and possibly less error prone than re-creating it each time.

On the other hand, [2] will still work if you switch to a different release of the same force field, while [1] in some situations might yield incorrect results.

Both approaches have their advantages and disadvantages. For proteins, [1] is probably better because of the reasons Rick gave (and the protein force field is pretty stable by now). For CGenFF molecules, [2] is probably preferable because the force field gets updated often.