R is an environment for
statistical analysis and visualization. As a
result it is well suited for the development and
analysis of QSAR/QSPR models as well as chemical datamining in general.
A package is available
that provides some utilites for handling ADAPT
generated data and file formats. Currently its
features include
Currently, the package does not carry out any Type I, Type II or Type III model development as defined in ADAPT and thus the ADAPT suite of programs is still required. However after the descriptor and set data has been read into R, they may be used as input to the various modeling and statistical functions available in R include linear regression (lm), generalized linear models (glm), neural networks (nnet), random forests (rf), recursive partitioning (rpart) etc.
Future PlansFuture plans for the package are to provide the functionality of genlin, annlin, gendes and qnetin so that Type I, II and III model development can be carried out within the R environment. Long term plans include interfacing the descriptor calculation routines with R - this will probably require a rewrite of the descriptor routines either as C code or R code though the latter would require the development of a full chemical representation subsystem (input of chemical formats, structure representation etc)Download & InstallationDownload the package hereTo install the package, from the command line do: R CMD INSTALL ADAPT_1.0.tar.gz. This will install the package in the system wide library directory. If you prefer to install it somewhere else do: R CMD INSTALL adapt_0.8.tar.gz -l /some/other/path |