The "research-type" questions are just taking what you know and applying it in a lateral form.
An example would be the lac operon, where you'll need to know the concept of it well, but on the actual exam, you might see a scientist doing an experiment on some novel gene system that just so coincidentally seems to be comparable to the lac operon.
On my exam, a year ago, I had two tough research-type molecular bio questions.
One asked about gene transfer techniques, where they wanted to know which step was required before proceeding, but literally they all made sense (i.e., you couldn't simply eliminate stupid answer choices), and it came down to actually understanding the step by step procedure for that process. I made an educated guess, but I still don't know whether I got it right.
The second question asked something about how to optimize gene acquisition via a plasmid, and then they had a chart listing three different plasmid components in the top row, and the answer choices below were basically different combinations (i.e., yes, no, or -) for whether you'd need that particular plasmid component.
The questions were tough because none of the Step1 resources drills that stuff. But getting lower yield questions on your exam is inevitable.
In the weeks following your exam, you're guaranteed to be aware of several questions you got wrong, but the research-type questions will always sit with you as an uncertainty because you likely won't remember them well enough to do the research on them later on.
