Functions to fit protective, oppoistional and coactive BRAID surfaces, and more specific flipped surfaces if necessary.
fitBraidFlipped(formula, data, flip, model, links = NULL, ...)fitProtectiveBraid_A(formula, data, ...)
fitProtectiveBraid_B(formula, data, ...)
fitOppositionalBraid_A(formula, data, ...)
fitOppositionalBraid_B(formula, data, ...)
fitCoactiveBraid_pure(formula, data, ...)
fitCoactiveBraid_partial(formula, data, ...)
A fit object of class braidrmflip. This structure contains the
exact same elements as an object of class braidrm (see braidrm() for
details) along with one additional element: flip, a character value
specifying the direction that the surface is flipped. The object's
coefficients and flip fields can be used to evaluate and invert the best
fit response surface using evalFlippedBraidModel() and
invertFlippedBraidModel().
Either an object of class formula such as would be provided
to a modeling function like stats::lm(), or a width-2 numeric array vector
of concentration pairs (including 0 or Inf). A formula should specify a
single output as a function of two inputs, eg. activity ~ conc1 + conc2.
If forumula is a symbolic formula, a data frame containing the
specified values. If formula is a numeric array of concentrations, a
numeric vector of response values, the same length as the number of rows of
formula.
String specifying the direction or directions of the surface's flip. Must be one of "A", "B", or "both".
Parameters model and links are used to specify which
variant of the BRAID model is fit to data. Model may be one of the
following character strings: "kappa1", "kappa2", or "kappa3" (see Details),
or a subset of the numbers 1 through 9 specifying which of the nine BRAID
response surface parameters is allowed to vary when fitting. links allows
the user to further specify constraints on the three BRAID maximal effect
parameters (see Details for more). If model is one of the supported
character strings, the parameter links will be ignored.
Additional parameters to be passed to braidrm
Though fitBraidFlipped offers the option of fitting any flipped BRAID
surface model specified by flip, this is not recommended, as the interplay
between flipping paramteers and parameter constraints becomes very confusing
very quickly. In nearly all cases, it is preferable to use one of the
pre-defined flipped fitting functions.
fitProtectiveBraid_A and fitProtectiveBraid_B fit "protective" surfaces
in which one drug has no effect in isolation, but attenuates or eliminates
the effect of the other. fitProtectiveBraid_A generates a surface in which
drug A is active and is attenuated by drug B; fitProtectiveBraid_B
generates the reverse.
fitOppositinalBraid_A and fitOppositionalBraid_B produce "oppositional"
surfaces in which a second drug produces an effect that is in the opposite
direction to the first drug, but which is then overwhelmed by the effect of
the first drug at higher concentrations. fitProtectiveBraid_A generates a
surface in which the maximal effect of drug A dominates a high concentrations,
fitProtectiveBraid_B generates the reverse. Note that the A and B in
the function names specify which compound's effect is dominant, not the
direction of the underlying flip; in actuality the surfaces generated by
fitProtectiveBraid_A are produced by flipping along the B axis.
fitCoactiveBraid_pure and fitCoactiveBraid_partial produce "coactive"
surfaces, in which both drugs have no or minimal effect in isolation, but
produce a pronounced effect when both are present. fitCoactiveBraid_pure
generates surfaces in which both drugs have no effect at all in isolation;
fitCoactiveBraid_partial generates surfaces in which either drug may have
a smaller partial effect in isolation.
fbfit1 <- fitProtectiveBraid_A(measure ~ concA + concB,
protectiveExample, getCIs=FALSE)
coef(fbfit1)
fbfit2 <- fitOppositionalBraid_A(measure ~ concA + concB,
oppositionalExample, getCIs=FALSE)
coef(fbfit2)
fbfit3 <- fitCoactiveBraid_pure(measure ~ concA + concB,
coactiveExample, getCIs=FALSE)
coef(fbfit3)
Run the code above in your browser using DataLab