r03ert0
commented
10 years ago Hi Nicolas,
You have to provide first of all the seed for making a coactivation map! This is an example command for saving the top tags:
./cmtool $dir/coincidences2013 -vol 6,24,23 -convert phir -out ~/Desktop/ -save topmesh
it will save to the Desktop a file called top100.0.txt with the 100 tags that correlate the most with the coactivation map with seed 6,24,23 (in voxel coordinates). This is the beginning of the resulting file:
R2 MeshCode 0.358088 Motor Skills 0.325081 Touch Perception 0.321242 Imitative Behavior 0.304931 Pitch Discrimination 0.304052 Imagination 0.301811 Psychophysics 0.299863 Sound Localization ...
The 1st column is the R2 (variance captured by the correlation), and the 2nd column is the MeSH name.
One important thing with cmtool is that the arguments are read sequentially, like a script. It is important then, for example, to define the output file (-out ~/Desktop/) BEFORE saving the results (the "-save topmesh" part), or to have defined the seed of the coactivation map (-vol 6,24,23) before converting it to correlations (-convert phir).
Now, instead of just getting the top 100 tags, you can also pull all the papers that constitute a specific map, with their tags. This is an example command:
./cmtool $dir/coincidences2013 -vol 6,24,23 -convert phir -out ~/Desktop/ -save articles 0.1 4
This is the result (first lines):
Peaks Percent ExpNumber PubMedID Reference MeshCodes Domains 1 12.5 1 15858160 "Abe N, 2006" F01.145.813.157;F03.300 1 12.5 3 18372290 "Abe N, 2008" F01.145.813.157;F02.463.785.626;F02.463.425.540;F02.463.425.540.706;F01.393.821
1 12.5 11 19015036 "Abel S, 2009" F02.463.188;F01.525.200.310 1 12.5 12 16487726 "Abler B, 2006" F02.463.425.234;F01.470;F01.658.370;F02.463.425.770.836 1 12.5 13 17947346 "Abutalebi J, 2008" F02.463.188 2 25 14 17947346 "Abutalebi J, 2008" F02.463.188 3 37.5 15 17947346 "Abutalebi J, 2008" F02.463.188 1 12.5 17 15670688 "Achim AM, 2005" F04.754.720.346;F02.463.425.540;F02.463.425.540.706 1 12.5 18 15670688 "Achim AM, 2005" F04.754.720.346;F02.463.425.540;F02.463.425.540.706 ...
The MeSH descriptors are organised as a tree. Each descriptor has a code and an equivalent name. By default, cmtool displays the codes. The codes start with a character, for example A for the codes related to anatomy, F for those related to psychiatry and Psychology, or C for diseases. By default, cmtool will display only the descriptors in the F category (check http://www.nlm.nih.gov/cgi/mesh/2014/MB_cgi). It is possible though to filter them otherwise using the "-meshfilter root", where you change "root" by the root of the codes you want to select. For example, within F, all mental disorders are under "F03". Then, if you wanted to have the top 100 mental disorders correlated with the coactivation map with seed 6,24,23, you should write:
./cmtool $dir/coincidences2013 -vol 6,24,23 -convert phir -out ~/Desktop/ -meshfilter F03 -save articles 0.1 4
which produces:
Peaks Percent ExpNumber PubMedID Reference MeshCodes Domains 1 12.5 1 15858160 "Abe N, 2006" F03.300 1 12.5 3 18372290 "Abe N, 2008" 1 12.5 11 19015036 "Abel S, 2009" 1 12.5 12 16487726 "Abler B, 2006" 1 12.5 13 17947346 "Abutalebi J, 2008" 2 25 14 17947346 "Abutalebi J, 2008" 3 37.5 15 17947346 "Abutalebi J, 2008" 1 12.5 17 15670688 "Achim AM, 2005" 1 12.5 18 15670688 "Achim AM, 2005" 1 12.5 20 17023179 "Addis DR, 2006" 2 25 21 17023179 "Addis DR, 2006" 1 12.5 22 19703572 "Adolph O, 2010" 1 12.5 23 19703572 "Adolph O, 2010" 1 12.5 26 20394829 "Agam Y, 2010" F03.550.325 ...
You can also choose to display names instead of codes using
./cmtool $dir/coincidences2013 -vol 6,24,23 -convert phir -out ~/Desktop/ -meshfilter F03 -meshdisplay names -save articles 0.1 4
which produces:
Peaks Percent ExpNumber PubMedID Reference MeshCodes Domains 1 12.5 1 15858160 "Abe N, 2006" Dissociative Disorders 1 12.5 3 18372290 "Abe N, 2008" 1 12.5 11 19015036 "Abel S, 2009" 1 12.5 12 16487726 "Abler B, 2006" 1 12.5 13 17947346 "Abutalebi J, 2008" 2 25 14 17947346 "Abutalebi J, 2008" 3 37.5 15 17947346 "Abutalebi J, 2008" 1 12.5 17 15670688 "Achim AM, 2005" 1 12.5 18 15670688 "Achim AM, 2005" 1 12.5 20 17023179 "Addis DR, 2006" 2 25 21 17023179 "Addis DR, 2006" 1 12.5 22 19703572 "Adolph O, 2010" 1 12.5 23 19703572 "Adolph O, 2010" 1 12.5 26 20394829 "Agam Y, 2010" Child Development Disorders, Pervasive 1 12.5 29 16337813 "Aichhorn M, 2006" 2 25 32 19439183 "Akitsuki Y, 2009" 1 12.5 34 19439183 "Akitsuki Y, 2009" 4 50 37 15269107 "Aleman A, 2005" 2 25 38 15269107 "Aleman A, 2005" 5 62.5 39 15238440 "Alkadhi H, 2005" 1 12.5 42 16023375 "Allen G, 2005" 1 12.5 44 19703570 "Allen P, 2010" Psychotic Disorders 1 12.5 45 19703570 "Allen P, 2010" Psychotic Disorders ...
In all the previous examples the coactivation map is that produced by the seed voxel at coordinates 6,23,24, but you can use any other of the methods to specify the seed: mni coordinates, an analyze volume, or a text file with MNI coordinates. If you have more than one seed, you can either get one file per seed, or make the average of all seeds (using the -average command after the -convert command)
hope this helps!
cheers, roberto
On Mon, Jan 20, 2014 at 12:47 AM, Nicolas Farrugia <notifications@github.com
wrote:
Hi,
when trying
./cmtool /Volumes/SubthalamicNucleus/Recherche/coactivationmap/cmap13 -tag affect -convert lr -out ~/Desktop/ -save map
I get a segfault...
Am I missing something ? Does the option -tag have to be combined with something else ?
Nico
— Reply to this email directly or view it on GitHubhttps://github.com/r03ert0/cmtool/issues/1 .
Hi,
when trying
./cmtool /Volumes/SubthalamicNucleus/Recherche/coactivationmap/cmap13 -tag affect -convert lr -out ~/Desktop/ -save map
I get a segfault...
Am I missing something ? Does the option -tag have to be combined with something else ?
Nico