README.md
In lxb: Fast LXB File Reader

Fast LXB file reader

This package provides functions to quickly read LXB parameter data. LXB is the format used by Luminex bead arrays and is based on the FCS v3.0 standard.

The package is available on CRAN so it can be installed like any other R package. For example, open R and type

packages.install("lxb")

Note that the functions in this package were written to run as fast as possible and with very specific LXB files in mind. It will not work with general files based on the FCS v3.0 standard and it must run on a little endian machine (e.g. Intel is ok, PowerPC is not). Adding support for more features of FCS v3.0 and other machines should be simple.

Here are some assumptions made:

The LXB file must be smaller than 100 Mb ($BEGINDATA and $ENDDATA are ignored)
The data must be integral, in list mode, and in little endian byte order ($DATATYPE = I, $MODE = L, $BYTEORD = 1,2,3,4)
At most 99 parameters are supported
Unicode in the text segment is not supported
Keys or values containing the separator character are not supported

## Typically your LXB files should be organized with one folder per
## plate, with each plate consisting of multiple wells and each well
## corresponding to one LXB file.  Assuming the folder 'plate1' contains
## all LXB files for the first plate, here is how to read all parameter
## data for plate 1 into a list of matrices (each list item is one well,
## each column in a matrix corresponds to one parameter):
y <- readLxb('plate1/*.lxb')

## It is now possible inspect individual wells, e.g. the dimensions of
## the first well are given by:
dim(y[[1]])

## .. and the names of the parameters for well 1 are given by:
colnames(y[[1]])

## If the LXB files have names like "XXXX_B1.lxb" (where 'B1' indicates
## that the LXB file corresponds to the first well on the second row),
## then it is also possible to index 'y' by the well name, e.g.:
dim(y$B1)

## .. which is the same as:
dim(y[[2]])

## You can see all well names (and the order of the wells in 'y') by
## typing:
names(y)