Format of the external variable file

Files with the external variables to be imported must be plain text ASCII files (.txt) containing the data as a matrix, where each row correspond to a compound and each column correspond to a variable. The first row should contain the labels (i.e. names) to be used for the variables. The number of rows, besides the first label row, must correspond to the number of molecule processed in the current session.


Each value must be separated by a tabulation character, that can be assigned by the user (the most commonly used is the TAB character).


The code for missing values can be assigned by the user, and it can be any alphanumeric code (common codes are “-999”, “n.a.”, “NaN”).


The importing procedure in Dragon graphical interface allows the user to decide to skip an arbitrary number of rows in the beginning of file, and to skip some columns. When using Dragon command line interface, please note that the user can not skip any row or column, so it is mandatory that the file contains exactly the matrix of data to be imported, with the first row containing variable labels.