Trying to run the unmodified cor.py tool from a current galaxy-dist we're getting this output using the sample input data:
% python /project/dbw-galaxy/DEVELOPMENT/tools/stats/cor.py /project/dbw-galaxy/DEVELOPMENT/test-data/cor.tabular /tmp/cor.out.ry4an 2,3 pearson Error: is.atomic(x) is not TRUE
which is happening in galaxy too. The line it's triggering is:
71 # Run correlation 72 try: 73 value = r.cor( array( matrix ), use="pairwise.complete.obs", method=method ) 74 except Exception, exc: 75 out.close() 76 stop_err("%s" %str( exc ))
and tracing through the execution the inputs look fine:
Breakpoint 1 at /home/msi/ry4an/cor.py:73 (Pdb) run Restarting ./cor.py with arguments: /project/dbw-galaxy/DEVELOPMENT/test-data/cor.tabular /tmp/cor.out.ry4an465 2,3 pearson > /home/msi/ry4an/cor.py(6)<module>() -> """ (Pdb) c > /home/msi/ry4an/cor.py(73)main() -> value = r.cor( array( matrix ), use="pairwise.complete.obs", method=method ) (Pdb) p matrix [[68.0, 4.0999999999999996], [71.0, 4.5999999999999996], [62.0, 3.7999999999999998], [75.0, 4.4000000000000004], [58.0, 3.2000000000000002], [60.0, 3.1000000000000001], [67.0, 3.7999999999999998], [68.0, 4.0999999999999996], [71.0, 4.2999999999999998], [69.0, 3.7000000000000002], [68.0, 3.5], [67.0, 3.2000000000000002], [63.0, 3.7000000000000002], [62.0, 3.2999999999999998], [60.0, 3.3999999999999999], [63.0, 4.0], [65.0, 4.0999999999999996], [67.0, 3.7999999999999998], [63.0, 3.3999999999999999], [61.0, 3.6000000000000001]] (Pdb) p array(matrix) array([[ 68. , 4.1], [ 71. , 4.6], [ 62. , 3.8], [ 75. , 4.4], [ 58. , 3.2], [ 60. , 3.1], [ 67. , 3.8], [ 68. , 4.1], [ 71. , 4.3], [ 69. , 3.7], [ 68. , 3.5], [ 67. , 3.2], [ 63. , 3.7], [ 62. , 3.3], [ 60. , 3.4], [ 63. , 4. ], [ 65. , 4.1], [ 67. , 3.8], [ 63. , 3.4], [ 61. , 3.6]])
The closest thing I could find ont the rpy list is:
http://www.mail-archive.com/rpy-list@lists.sourceforge.net/msg02612.html
which hints at an input conversion problem in rpy.
We've got rpy 1.0.3 and R version 2.10.1 on this galaxy instance.
Has anyone else seen this, or does anyone have a suggestion as to how the tool might be tweaked to bypass it?
Thanks,