All well and good, until it's time to start looking at what restrictions are "necessary" and which are "unnecessary".
The scientific method restricts submissible evidence to being what is independently verifiable. This is done in pursuit of trying to insure that the research is done objectively. Submitting that some disciplines or subjects should not be subject to that restriction means they will not be held to that standard of objectivity.
What's the criteria for deciding which disciplines and subjects need to conform to that standard, and which don't? Is molecular biology a "broad" subject, or a "narrow" subject?
Within the field of molecular biology, an investigation might be very narrowly defined, e.g. designing an antigen to provoke an antibody to study it - or it may be broadly defined, e.g. rise of syntactical autonomy (Rocha, Pattee et al.)
I would imagine that the more narrowly defined investigations would have established protocols which entail many of the usual restrictions.