<My $0.02: The only thing I can think of immediately is familiarity with concepts like signal-to-noise ratio and open loops (hardware) plus GIGO and perhaps fuzzy logic (software) ;^)>
imnot6,
Engineering is a fine profession and you make some good points. However, there is a fundamental difference between biology and engineering. One of the bigwigs at INTEL had prostate cancer and was talking about using the model of computer chip design and advance where a doubling of speed is about every 18 months, and applying it to cancer research. It is not possible. In ChIP design or engineering for construction, they know the precise physical properties of materials and forces the structure must deal with, so a systematic and logical approach is possible. They aren't going to find out that steel or concrete has some fundamental new properties they never dreamed of, which alter how the building or oil rig will respond to winds and pressure. Yes you can screw up the manufacturing of steel and concrete or neglect to factor in some forces, yes, but there isn't a fundamental uncertainty about basic properties of steel or concrete.
This happens all the time in biology and cancer research. Our knowledge is still relatively limited, incomplete and often wrong. This means the prevailing dogma is subjected to small revisions in our views, and yes, even to "paragdim shifts" that fundamentally alter our views. I think that two of the most recent paradigm shifts were the discovery of the RNAi (or siRNA) pathways, which include its relatives microRNAs. These not only provided tools for genetic examination of vertebrate cells, but also illustrated how complex the processes of transcription and chromatin remodeling were, and how they can be inter-related. The second was the ability to reprogram fully differentiated cells into stem cell-like states and the subsequent discovery that this can be accomplished by over-expressing only a few genes.
For cancer research , one uses tissue culture cells and animal models to identify promising candidates for clinical trials. However, both are crude approximations and candidates often fail in humans trials. You inhibit some growth pathway, but some other redundant pathway kicks in to alter the results, and generate resistance. Such pathways may not be stimulated in culture cells, or don't operate quite the same way in animal models as in humans. You get the picture.
AI
