In fact, there are some general rules:
morphine QSAR
1. Enantiomeric dextrorotatory morphine is devoid of analgesic activity.
2. A trans-ring junction between rings Band C drastically reduces analgesic potency.
3. A phenolic hydroxyl at C-3 is important for analgesia, but not necessary. The 3,6-deoxy-derivative is equipotent to morphine but may represent metabolic hydroxylation.
4. Other substituents on the A-ring decrease or eliminate activity.
5. Substituents at C-l0 in the B-ring maintain or decrease morphinomimetic
activity.
6. Some of the chemical features of the Coring of morphine are relatively noncrucial: removal of the double bond (7,9) or the alcohol at C-6. Numerous other chemical modifications are also compatible with enhanced analgesic activity: a C-5 methyl group, oxygenation or alkylation at C-6, phenylalkyl substitution at 7-beta, and short alkyl substitution at 8-beta.
7. Hydroxyl substitution at the BC-ring junction (C-14), as well as its acyl derivatives, strongly enhances analgesia. Other substituents eliminate activity.
8. The methyl group on the amine is not optimal. Replacement by arylalkyl or functionalized arylalkyl groups increases morphinomimetic activity severalfold. Replacement by N-lower alkyl (propyl, allyl, cyclopropylmethylene) produces morphine antagonists.
9. Addition of a sixth ring by Diels-Alder additions to thebaine can increase either morphine agonist or antagonist activity by at least four orders of magnitude.
Diels-Alder adducts
1. The substitution on the carbinol group at C-7 is important: C-19 tertiary alcohols have a higher analgesic potency than secondary ones; however, in cases where both C7-alpha and C-7beta 3 epimers have been evaluated, only slight differences in potency have been observed in the pairs.
2. Highest activities have been found when a moderate disparity in size between the two groups on the C-19 alcohol, R" and R', exists. One substituent should be small, that is, H or CH3. Maximal analgesic activity is then found when the second substituent is a 3- to 5- hydrocarbon chain; further lengthening results in a steady decrease in activity.
3. The diastereomers of unsymmetrical tertiary alcohols can show markedly different analgesic properties, the 19R configuration being the more potent morphine agonist.
4. The comparable 6, 14-endoetheno and 6,14-endoethano analogs show only marginal differences in analgesic agonist or antagonist potency.
5. The oripavine derivatives (C-3 hydroxyl) or their C-3 acetylated analogs are more potent analgesics than the thebaine bases (C-3 methoxy).
6. The piperidino nitrogen should be tertiary, the secondary amines being less active as agonists. Increasing the size of the nitrogen substituent beyond methyl steadily decreases analgesic activity. Some N-allyl and N-cyclopropylmethyl thebaine and oripavine derivatives
of tertiary alcohols are potent antagonists.
7. The piperidine and ether rings should remain intact. Substitution on the piperidine ring near the basic nitrogen creates potent antitussives.
This information is almost 30 years old so if anyone has more modern data, please share. I seem to remember that the benzopyrene analogs were pretty strong when substituted for the A,B & C rings in the morphine molecule (x60 if I remember correctly). They also miss quite a lot of possible C-14 substituants (semaphorine for example).