Correlation by lines
This software takes tables as input and computes the Pearson correlation coefficient between all combinations of rows.
Manual :
The software was written in GO and compiled on linux 64 bits with 15 threads.
1- install Perl free programming language and GNU parallel.
2- unzip the software
3- copy your n tables in csv (TAB delimitated by default) files in the “data” directory.
The tables must have at least 3 rows.
- first line of the table : columns names
- first column : rows names
- separator : TAB
- decimal separator : '.'
- the software accept 'NA' and 'NaN' in the table
4- Edit the parallel configuration file to set the max number of CPUs = max number of tables processed in parallel.
the executable file uses 8 threads per table. So if you set nb_cpu=4 the sofware will use 4x8=32 threads.
5- execute the software by the command : perl corr_by_lines-0.8.pl
6- processed file is in the “result” directory.
Caution : If you process a very large table, the results table will be too large to be opened in a spreadsheet. To open the table in R, don't forget to skip the first line using : read.csv(file="file.csv", skip=1, header=F, sep = '\t')
7- you can use the command : perl parallel_format_table-0.1.pl to obtain a correlation table with means and SD.
8- formatted table is in the table directory