Talk:Structure–activity relationship

SAR and SAR paradox: I'm removing the somewhat alien reference to the No Free Lunch (NFL) theorem. Regardless of what NFL says about black-box search in general, it clearly doesn't apply in such a regular environment as inducing structure-activity relationships. The underlying relationships - for biological activity as a function of which functional groups are present in what combination - are physical, hence continuous. Continuous non-fractal functions don't follow the NFL theorem. Wikisteff (talk) 14:32, 16 January 2008 (UTC)

QSAR vs. visual or modelled SAR in engineering new test compunds
Utilization of modeling software and colorization to compare like-looking results, in conjunction with reports from human volunteers or animal observers can also be used to find trends in SAR, and this approach is every bit as valid as a purely matrix-oriented QSAR approach; use of matrixes often provides only a little insight into non-linear phenomenon in general.

Most human beings are very well equipped to make excellent SAR guesses; one need only watch a toddler learn to fix blocks into a box to realize how fundamentally simple of a problem this actually is, and how scientists have really gone out of their way to make appreciating SAR in biochemistry something only approachable from understanding gained in the ivory tower of academia. But it doesn't have to be this way; people just need better software, tools, and simulations for visualizing and appreciating SAR at a scale compatible with the human experience.

Seriously, as totally valid as the SAR approach to understanding the unmapped (or unfolded) receptor, or even the spectrum of side effects produced by any given drug (that is, how it fits into multiple holes) relative to its near analogs .. the quantitative approaches seem to me to be more useful after the fact when it comes time to publish a paper in a major journal. Math impresses. But in a field that very easily has dozens of times more left to discover than what we already know, a math-based approach like QSAR is probably less likely to result in discovery than a more visual or even physical implementation of SAR modeling.Zaphraud (talk) 19:26, 17 April 2011 (UTC)

Encourage strong scholarly attention to this article
As foundational as these concepts are -- fundamental to QSAR, and underlying all pharmacologic and drug discovery practice -- this article is woefully lacking in substance and breadth. Please, readers with expertise, take time and make this into the foundation stone it needs to be. Prof D.   Meduban (talk) 04:25, 30 August 2011 (UTC)

SAR Paradox?
Why is this refered to as a paradox? If your hypothesis is true some of the time but not all of the time, that's a crappy hypothesis, not a paradox! Simon G. 128.101.196.178 (talk) 23:27, 3 May 2013 (UTC)