pd/npd as an output as well as
npd/npdenlmixr2 “saem” fit from another R
session, nlmixr2 will no longer crash with
fit$objfNPDE was identical to NPD even with
correlated models, this was fixed (prior output was actually
NPDE).Added ui$getSplitMuModel which is used in
babelmixr2 and will be used in the refined stepwise
covariate selection of nlmixr2extra
Added work-around to remove
_nlmixr2est_RcppExport_registerCCallable since the
registering of C callable are handled manually at the moment.
Use .zeros() for the matrices in armadillo in
addition to relying on calloc to give zero
matrices.
Fixed one uninitialized object
Fix for augPred so it works on population only
models
nlme no longer sets options to treat all covariates
as non mu-referenced covariates, but directly calls a function that can
turn on or off the mu-reference covariate selection.
vpcSim now tries to simulate IDs that didn’t
simulate correctly (with a warning)
Export nmObjHandleControlObject
nlmixr2est contains the estimation functions within
nlmixr2.
Remove lower level foceiFit function. Focei, foce,
fo, foi, and posthoc now directly takes rxode2 ui objects
New error types are supported in focei including mixing theta and etas in residual errors and different types of proportional errors
Different types of additive and proportional errors can be used
for each endpoint using + combined1() or
+ combined2() otherwise it takes the supplied
addProp option to figure out which type of combined model
is run (by default combined2())
Focei model cache is now named focei-md5Digest.qs
and uses qs compression/saving/loading.
foceiControl() aligned between other
methods.
foceiControl(adjLik=TRUE) uses the NONMEM-style
objective function throughout. foceiControl(adjLik=FALSE)
uses the adjusted objective function throughout, and adjusts it back to
the NONMEM objective function.
Lag time and other between subject variability differences no longer calculate an ideal relative step size, but an absolute step size when using Gill differences (default)
Objective function checks for infinite/NaN/NA values for the entire solving space and ensures no overflow occurs when calculating the inner hessian
mu referencing is no longer required for saem;
Internally non mu-referenced values are converted to mu referenced
values and the converted back when calculating the nlmixr2
object.
nlmixr2 forced the parameter ordering to (1)
population effects,
nlmixr2
sees the parameters. Since this is based on a random number generator,
the optimization trajectory will be different and have different results
than nlmixrComponents of omega can now be fixed.
Residual error components can also be fixed.
When optimizing only one residual value, nlmixr2’s saem uses
nlm from R, which is more efficient than the nealder-meade
method.
Lower level saem functions (like
configsaem()) are not exported because they are
increasingly difficult to use and convert to something standard; a few
methods (like print, summary etc) are
maintained to view the lower level object and for debugging it.
Parameter history and print-out no longer includes fixed parameters.
The model to calculate the residuals more closely matches the model used for estimation to remove small rounding differences that may occur in the models.
Different types of additive and proportional errors can be used
for each endpoint using + combined1() or
+ combined2() otherwise it takes the supplied
addProp option to figure out which type of combined model
is run (by default combined2())
Parameter history and printout now uses standard deviation for additive only components, matching the estimation of the components.
rxode2 solving options are now saved in the
rxControl part of the saemControl(). That is
saemControl(rxControl=rxControl(...)); This fixes any
conflicting option names as well as allowing alignment between the
control structures in focei, nlme and
saem
saemControl() aligned between other
methods.
nlme has been completely rewritten to directly run
from the rxode2 UI
nlme always tries to use mu-referencing (when
available)
Internally nlme now uses parallel processing for
solving so it should be faster.
nlmixr2NlmeControl() (which will overwrite
nlmeControl()) documents and adds more options to
nlme. Also aligned with other methods.
weights, fixed, random can
be specified in nlmixr2NlmeControl(). If so, then the
nlme object will be returned.
returnNlme is a new option that will return the
nlme object instead of the traditional nlme
object.
nlme_ode and lme_lin_cmpt are both
removed.
rxode2 solving options are now saved in the
rxControl part of the saemControl(). That is
nlmeControl(rxControl=rxControl(...)); This fixes any
conflicting option names as well as allowing alignment between the
control structures in focei, nlme and
saem
With saem, the nlmixr2 function now saves/compresses
the phiM information. This means the gaussian and laplacian
likelihoods can be calculated when you save the nlmixr object and then
restore it later.
The nlmixr2 object compresses infrequently used and removes many
unneeded objects. Even with compression, the saem objects
are often a bit bigger since they include the large phiM
object.
nlmixr2 now supports non-mu referenced ETAs in the
fit$parFixed and fit$parFixedDf
nlmixr2 interface changed to use rxode2
UI
keep and drop are added to
tableControl to influence the end data-frame
$simInfo uses a quoted expression for
$rx instead of a string
$simInfo$sigma is a diagonal matrix since now the
normal simulation is controlled by the variability modeled as a
population value.
nlmixr2 now allows etas that have initial omega
estimates of zero to be dropped from the model (instead of issuing an
error about a non-positive definite $omega matrix)
addNpde(fit, table=tableControl(nsim=500))vpc function rewritten and split out to
vpcSim() and vpcPlot() (which is a replacement
for vpc()).
There were too many mismatches between vpc::vpc and
nlmixr::vpc which caused inconsistencies in code based on
load order of vpc and nlmixr. This way both
coexist, and you can use the vpc simulation for other
packages more easily (like ggPMX) without creating or
summarizing data since ggPMX has its own methods for
summarizing and creating plots.
VPC now directly uses rxode2::rxSolve
augPred() has been written to use the new fit
object.
nlmixr2AugPred was changed to
nlmixr2AugPredSolve()
augPred uses the new interface and supports multiple
endpoints. The endpoint name is now always on the
plot(augPred(fit)).
fit$est,
and now getFitMethod(fit) simply returns
fit$estMany methods lower level utility functions have been deleted.
nmDocx, nmLst and nmSave
have been removed.
function 'rx_0ba247452048de33b1ffb8af516714fc__calc_lhs' not provided by package 'rx_0ba247452048de33b1ffb8af516714fc_'
would cause the estimation to stop. Now rxode2::rxClean()
is run when this occurs.focei subject initialization, see #566Fix for nlmixrSim CMT to have a factor that matches
the RxODE definition (issue #501)
Give instructions on how to reinstall nlmixr if it is linked to a
different version of RxODE. (#555)
Now inform which parameters are near the boundary (#544)
The saem estimation routine will now increase the
tolerance when ODE solving is difficult; This can be controlled with
odeRecalcFactors and maxOdeRecalc. This is
similar to the handling that focei already uses.
For focei family estimation methods:
If the inner problem couldn’t solve the ODE using the forward sensitivities, try using numerical differences to approximate the derivatives needed for the focei problem. A warning will be issued when this occurs. This requires RxODE 1.1.0 that always generates the finite difference prediction model. If RxODE is an earlier version, only apply this when the finite differences are supplied to nlmixr. This occurs when there are ETAs on the dose based events like duration, lag time, bioavailability etc.
If eta nudge is non-zero, when resetting an ETA estimate, try the zero estimate first, and then the nudged locations.
When there is an ODE system for an individual that cannot be
optimized in the inner problem, adjust that individual’s objective
function by 100 points. This can be controlled by
foceiControl(badSolveObjfAdj=100)
Theta reset now will now make sure the parameter is estimated and between the proper bounds before resetting.
$simInfo non longer tries to generate the covariance
step, and will simply have a $simInfo$thetaMat entry of
NULL if the covariance step was unsuccessful.
With vpc() if the cmt conversion isn’t working
correctly, fall back to compartment numbers.
Take out symbol stripping based on CRAN policies
Fall back gracefully when rbind doesn’t work in
parameter histories.
Correctly print out the number of compartments based on the new
RxODE linCmt() that was updated to support
solved systems in focei. (Reported by Bill Denney #537).
Use strict headers since Rcpp now is moving toward strict
headers. Also changed all the Calloc to
R_Calloc, Free to R_Free, and
DOUBLE_EPS to DBL_EPSILON.
gnlmm no longer imports the data.frame to an RxODE
event table. This should speed up the routine slightly and (more
importantly) make it easier to specify time varying covariates.
Now can use the following for combined error models:
foceiControl(addProp=1)
foceiControl(addProp=2) saemControl(addProp=1)
saemControl(addProp=2)
Bug-fix for SAEM add+prop and other error models that are optimized with nelder mead simplex (#503)
Bug-fix for more complex SAEM models that were not parsing and running. (Issue #502, #501)
Issue the “NaN in prediction” once per SAEM problem (#500)
Detection of initial conditions was rewritten to enable
additional features in the initial conditions (#322). The most important
user-facing change is that now arbitrary R expressions can be used when
setting initial conditions such as tvCL <- log(c(2,3,4))
(#253) instead of simply tvCL <- log(3)
The function as.nlmixrBounds() now supports adding the columns that are missing into the input data.frame.
omega definitions can be correlation matrices (#338)
Can specify keep= and drop= in the
nlmixr function to keep and drop columns in nlmixr output. Can also
specify control=list(keep=,drop=) or
nlmixr(...,keep=,drop=) to keep/drop columns
(#260)
focei changes:Uses RxODE to re-arrange the problem so it does not include
if/else in the model (ie. un-branched code). This allows
sensitivities to be calculated in one pass saving time for multiple
endpoint models and models with if/else in them.
linCmt() now uses solved systems instead of
translating to ODEs.
RxODE/stan’s math headers to
calculate the sensitivities of the super-positioned
linCmt() solutions.advan solutions and hence supports
support time-varying covariates.focei now supports censoring in the same way
monolix does, with cens and limit
columns
focei now allows etas on dose-related
modeled events like alag, f, etc by finite
difference sensitivities.
focei now supports 2 combined additive +
proportional error models;
combined1:
trans(y) = trans(f) + (a+b*f^c)*errcombined2:
trans(y) = trans(f) + sqrt(a^2+b^2*f^(2c))*errfocei etaNudge parameters were changed
to use quadrature points covering 95% percent of a standard
normal.
With zero gradients, Gill differences are recomputed to try to find a non-zero gradient.
Now when running if a zero gradient is detected, reset the
problem (theta reset) and re-estimated with
outerOpt="bobyqa"
Now when running a model where the last objective function is not the minimum objective function, issue a warning and skip the covariance step. (See Issue #403)
focei proportional and power models are more
tolerant of 0 predictions in your data
saem fits now gracefully fall back to the
focei likelihood when they support files are no longer on
the loaded disk
saem phi pile is now saved in the
RxODE::rxTempDir() which can be customized to allow the
phi file to remain after R has exited
saem fits now can add in fo,
foce and focei likelihood
saem fits now use liblsoda by default
and are multi-threaded when running (controlled by
RxODE)
saem now supports time-varying covariates (like
clock-time)
saem now supports 2 combined additive + proportional
error models:
combined1:
trans(y) = trans(f) + (a+b*f^c)*errcombined2:
trans(y) = trans(f) + sqrt(a^2+b^2*f^(2c))*errsaem proportional and power models are more tolerant
of 0 predictions in your data
saem now supports censoring a similar way as
monolix does, with cens and limit
columns
The default of saem additive + proportional error
has been switched to combined2, which was the
focei default, but you can change this back with
saemControl(addProp="combined2"). The table results will
likely be different because in the last release the saem
calculated combined1 and then used these coefficients in
the combined2 focei problem.
nlme will now support 2 combined additive +
proportional error models (if the patched version of nlme is used)
combined1: y = f + (a+b*f)*errcombined2:
y = f + sqrt(a^2+b^2*f^2)*errnlmeControl(addProp="combined1") to use
the combined1 type of error modelbootstrapFit now calculates the bootstrap confidence
bands and (optionally) will compare with the theoretical chi-squared
distribution to help assess their adequacy.
covarSearchAuto now allows automatic
forward/backward covariate selection
Added auto-completion of nlmixr object properties
accessed by $. This works for major editors including
Rstudio, ESS, and Base R itself.
Changed the way that Rstudio notebooks display
nlmixr objects; It should be more legible in
Rstudio.
Graphics have been revamped to show censoring (including adding
ggplot stat/geom geom_cens) as well as use
RxODE’s ggplot theme (rxTheme()). Additionally
time after dose is calculated as tad for all
nlmixr models
Tables generation has been refactored; npde uses the
arma and RxODE random number generators which
may change results. Also the default of ties=TRUE has been
changed to ties=FALSE. npde calculations have
been threaded with OpenMP to speed up the calculation as
well. This refactoring was required to have the dv
imputation between cwres and npde use the same
method. The npde option now calculates the decorrelated
npd as well, (which is the recommended weighted residual;
see Nguyen 2017)
saem and focei additive +
proportional error models, so saem
additive+proportional outputs will be different using the
correct focei methodNote this includes all the RxODE changes including dropping python.