FreeSurfer Longitudinal Analysis (implementacion en detritus)

https://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessing

https://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalStatistics

Tengo dos puntos del proyecto FACEHBI asi que puedo hacer un proyecto longitudinal (apenas). A tener en cuenta,

En cada punto temporal del proyecto un sujeto recibe un ID de procesamiento distinto, esta ID corresponde al sujeto dentro del pool en ese punto de tiempo. Asi que hay que cruzar los IDs de procesamiento con los IDs de proyecto que si son los mismos a lo largo del tiempo. Esto es bastante sencillo pero hay que hacerlo.
Para el analisis longitudinal hay que crear un proyecto distinto y crear IDs nuevos.Este proyecto sera especial ya que no hace falta ningun input mas alla de la correspondencia entre los puntos temporales. Todaslas imagenes estan ya procesadas. Solo hay que reprocesar punto temporal respecto a un template del sujeto que se hace con ellos.

Asi que tenemos algo asi,

[osotolongo@detritus facehbi_long_v02]$ head match02.csv
Psubject_v0,SubjID_v0,SubjID_v2
F001,facehbi_0001,f2cehbi_0001
F002,facehbi_0002,f2cehbi_0002
F003,facehbi_0003,f2cehbi_0003
F005,facehbi_0005,f2cehbi_0004
F006,facehbi_0006,f2cehbi_0005
F007,facehbi_0007,f2cehbi_0006
F008,facehbi_0008,f2cehbi_0007
F009,facehbi_0009,f2cehbi_0008
F010,facehbi_0010,f2cehbi_0009

Voy a probar con uno solo, despues ya veremos la logistica.

Single subject example

[osotolongo@detritus facehbi_long_v02]$ recon-all -base F001 -tp facehbi_0001 -tp f2cehbi_0001 -all
[osotolongo@detritus facehbi_long_v02]$ recon-all -long facehbi_0001 F001 -all
[osotolongo@detritus facehbi_long_v02]$ recon-all -long f2cehbi_0001 F001 -all

programando el grupo

Lo primero que necesito es algo un archivo que enlace los distintos puntos temporales,

[osotolongo@detritus facehbi_long_v02]$ head flong.csv
flong_subject;SubjID_v0;SubjID_v2
flong_0001;facehbi_0001;f2cehbi_0001
flong_0002;facehbi_0002;f2cehbi_0002
flong_0003;facehbi_0003;f2cehbi_0003
flong_0005;facehbi_0005;f2cehbi_0004
flong_0006;facehbi_0006;f2cehbi_0005
flong_0007;facehbi_0007;f2cehbi_0006
flong_0008;facehbi_0008;f2cehbi_0007
flong_0009;facehbi_0009;f2cehbi_0008
flong_0010;facehbi_0010;f2cehbi_0009

Despues se hace un script para lanzar todo esto en el cluster

plong.pl

#!/usr/bin/perl
 
use strict; use warnings;
use File::Basename qw(basename);
use Data::Dump qw(dump);
 
use NEURO4 qw(populate check_subj load_project print_help check_or_make get_subjects get_list);
 
my $ifile;
 
@ARGV = ("-h") unless @ARGV;
while (@ARGV and $ARGV[0] =~ /^-/) {
    $_ = shift;
    last if /^--$/;
    if (/^-i/) { $ifile = shift; chomp($ifile);}
    if (/^-h/) { print_help $ENV{'PIPEDIR'}.'/doc/longitudinal.hlp'; exit;}
}
 
my $study = shift;
unless ($ifile && $study) { print_help $ENV{'PIPEDIR'}.'/doc/longitudinal.hlp'; exit;}
my $debug = 1;
my %lsubjects;
 
open ILF, "<$ifile" or die "Could not open input data file\n";
while(<ILF>){
        my @long_list = split /;/, $_;
        my $lindex = $long_list[0];
        splice(@long_list,0,1);
        $lsubjects{$lindex} = [ @long_list ];
}
close ILF;
my %std = load_project($study);
my $data_dir=$std{'DATA'};
#open debug file
my $logfile = "$std{'DATA'}/.debug.controlled.log";
$debug ? open DBG, ">$logfile" :0;
#open slurm file
my $outdir = "$std{'DATA'}/slurm";
check_or_make($outdir);
foreach my $lsubject (sort keys %lsubjects){
        my $sub_check_dir = $ENV{'SUBJECTS_DIR'}.'/'.$lsubjects{$lsubject}[0];
        if( -e $sub_check_dir && -d $sub_check_dir){
                my $subol = join ' -tp ', @{$lsubjects{$lsubject}};
                chomp $subol;
                my $order = "recon-all -base ".$lsubject." -tp ".$subol." -all";
                my $orderfile = $outdir.'/'.$lsubject.'_fs_long_base.sh';
                open ORD, ">$orderfile";
                print ORD '#!/bin/bash'."\n";
                print ORD '#SBATCH -J fs_long_'.$study."\n";
                print ORD '#SBATCH --time=48:0:0'."\n"; #si no ha terminado en X horas matalo
                print ORD '#SBATCH --mail-type=FAIL,TIME_LIMIT,STAGE_OUT'."\n"; #no quieres que te mande email de todo
                print ORD '#SBATCH --mail-user='."$ENV{'USER'}\n";
                print ORD '#SBATCH -p fast'."\n";
                print ORD '#SBATCH -c 4'."\n";
                print ORD '#SBATCH --mem-per-cpu=4G'."\n";
                print ORD '#SBATCH -o '.$outdir.'/fs_long_base-slurm-'.$lsubject.'-%j'."\n";
                print ORD "srun $order\n";
                close ORD;
                my $jobid = `sbatch $orderfile`;
                $jobid = ( split ' ', $jobid )[ -1 ];
                foreach my $ind_subject (sort @{$lsubjects{$lsubject}}){
                        chomp $ind_subject;
                        my $order = "recon-all -long ".$ind_subject." ".$lsubject." -all";
                        my $orderfile = $outdir.'/'.$lsubject.'_'.$ind_subject.'_fs_long_base.sh';
                        open IORD, ">$orderfile";
                        print IORD '#!/bin/bash'."\n";
                        print IORD '#SBATCH -J fs_long_'.$study."\n";
                        print IORD '#SBATCH --time=48:0:0'."\n"; #si no ha terminado en X horas matalo
                        print IORD '#SBATCH --mail-type=FAIL,TIME_LIMIT,STAGE_OUT'."\n"; #no quieres que te mande email de todo
                        print IORD '#SBATCH --mail-user='."$ENV{'USER'}\n";
                        print IORD '#SBATCH -p fast'."\n";
                        print IORD '#SBATCH -c 4'."\n";
                        print IORD '#SBATCH --mem-per-cpu=4G'."\n";
                        print IORD '#SBATCH -o '.$outdir.'/fs_long_ind-slurm-'.$lsubject.'_'.$ind_subject.'-%j'."\n";
                        print IORD "srun $order\n";
                        close IORD;
                        system("sbatch --dependency=afterok:$jobid $orderfile");
                }
                sleep(2);
        }
}
 
$debug ? close DBG:0;
my $orderfile = $outdir.'/fs_long_end.sh';
open ORD, ">$orderfile";
print ORD '#!/bin/bash'."\n";
print ORD '#SBATCH -J fs_long_'.$study."\n";
print ORD '#SBATCH --mail-type=END'."\n"; #email cuando termine
print ORD '#SBATCH --mail-user='."$ENV{'USER'}\n";
print ORD '#SBATCH -o '.$outdir.'/fs_long_end-%j'."\n";
print ORD ":\n";
close ORD;
my $order = 'sbatch --dependency=singleton '.$orderfile;
exec($order);

Creo un proyecto fantasma,

[osotolongo@detritus lfacehbi]$ make_proj lfacehbi /nas/corachan/facehbi

y lanzo el longitudinal,

[osotolongo@detritus lfacehbi]$ plong.pl -i flong.csv lfacehbi

Sacando las tablas

Probando con una tabla, aseg, por ejemplo.

(https://surfer.nmr.mgh.harvard.edu/fswiki/freesurferstats2table?action=fullsearch&context=180&value=stats2table&titlesearch=Titles)

[osotolongo@detritus lfacehbi]$ asegstats2table --subjects `ls -d /nas/data/subjects/facehbi_*.long.flong_* | awk -F"/" '{print $5}' | sed ':a;N;$!ba;s/\n/ /g'` --meas volume --skip --statsfile aseg.stats --all-segs --tablefile facehbi_aseg_stats.txt
SUBJECTS_DIR : /nas/data/subjects
Parsing the .stats files
Building the table..
Writing the table to facehbi_aseg_stats.txt
[osotolongo@detritus lfacehbi]$ asegstats2table --subjects `ls -d /nas/data/subjects/f2cehbi_*.long.flong_* | awk -F"/" '{print $5}' | sed ':a;N;$!ba;s/\n/ /g'` --meas volume --skip --statsfile aseg.stats --all-segs --tablefile f2cehbi_aseg_stats.txt
SUBJECTS_DIR : /nas/data/subjects
Parsing the .stats files
Building the table..
Writing the table to f2cehbi_aseg_stats.txt
[osotolongo@detritus lfacehbi]$ sed 's/Measure:volume/Visit, Subject/;s/facehbi_.*\.long\./0, /' facehbi_aseg_stats.csv > facehbi_aseg_stats_clear.csv
[osotolongo@detritus lfacehbi]$ sed 's/Measure:volume/Visit, Subject/;s/f2cehbi_.*\.long\./2, /' f2cehbi_aseg_stats.csv > f2cehbi_aseg_stats_clear.csv
[osotolongo@detritus lfacehbi]$ tail -n +2 f2cehbi_aseg_stats_clear.csv > f2cehbi_aseg_stats_clear_nohead.csv
[osotolongo@detritus lfacehbi]$ cat facehbi_aseg_stats_clear.csv f2cehbi_aseg_stats_clear_nohead.csv
[osotolongo@detritus lfacehbi]$ grep flong_0001 facehbi_long_aseg_stats.csv
0, flong_0001, 11069.1, 335.9, 17456.1, 55694.4, 7314.9, 3132.9, 5000.3, 2141.7, 1048.7, 1651.0, 24749.2, 3472.6, 1279.8, 1564.1, 671.2, 3900.3, 159.1, 1318.8, 12173.8, 312.2, 16618.2, 53669.2, 7055.8, 3722.3, 5238.6, 2174.4, 3642.3, 1624.8, 657.0, 3968.7, 201.4, 1371.7, 0.0, 3905.4, 0.0, 0.0, 18.6, 0.0, 0.0, 290.4, 1108.3, 772.0, 760.4, 463.8, 1146.2, 1264641.0, 1234001.0, 1233494.22632, 239473.467727, 242179.420796, 481652.888523, 273169.386984, 275225.950812, 548395.337795, 56835.0, 651931.888523, 1118979.22632, 1092494.22632, 1087419.0, 1719786.0, 0.725663, 0.986829, 1742738.86176
2, flong_0001, 12198.3, 421.5, 17554.5, 54551.3, 6946.3, 2947.5, 4946.3, 2109.3, 1106.2, 1746.9, 24346.9, 3394.6, 1203.1, 1573.0, 653.2, 3740.2, 161.3, 1289.8, 13200.3, 365.7, 16464.4, 52906.5, 6795.9, 3625.8, 5045.0, 2093.4, 3558.4, 1529.8, 640.9, 3838.2, 230.3, 1391.3, 0.0, 4726.2, 0.0, 0.0, 34.1, 0.0, 0.0, 305.1, 1095.1, 681.4, 711.8, 459.0, 1099.7, 1225467.0, 1192355.0, 1192475.46698, 225610.010964, 226004.976163, 451614.987126, 275230.775936, 265861.703922, 541092.479858, 54917.0, 618316.987126, 1082197.46698, 1053410.46698, 1047595.0, 1718068.0, 0.703185, 0.985844, 1742738.86176

Me voy a R ahora y a ver si puedo sacar un ejemplo, (ver How to correct by Intracranial Volume (ICV))

aseg <- read.csv("facehbi_long_aseg_stats.csv", header = TRUE, sep=",")
aseg$Hippocampus = 0.5*(aseg$Left.Hippocampus + aseg$Right.Hippocampus)
a <- lm(aseg$Hippocampus ~ aseg$EstimatedTotalIntraCranialVol)
b = a$coefficients[[2]]
aseg$adj_Hippocampus = aseg$Hippocampus - b*(aseg$EstimatedTotalIntraCranialVol -mean(aseg$EstimatedTotalIntraCranialVol, na.rm=TRUE))
interaction.plot(aseg$Visit, aseg$Subject, aseg$adj_Hippocampus, legend=F, col=c(1:200), xlab="Visit", ylab="adjusted Hippocampus")

Statistics

Qdec Table

Tengo que sacar una tabla con los años entre cada experimento MRI. Primero saco las echas de cada experimento,

[osotolongo@detritus lfacehbi]$ for x in `ls /nas/corachan/facehbi`; do f=$(find /nas/corachan/facehbi/${x}/ -type f | head -n 1); d=$(dckey ${f} -k "AcquisitionDate" 2>&1); echo "${x}, ${d}"; done > v0_dates.csv
[osotolongo@detritus lfacehbi]$ for x in `ls /nas/corachan/facehbi_2`; do f=$(find /nas/corachan/facehbi_2/${x}/ -type f | head -n 1); d=$(dckey ${f} -k "AcquisitionDate" 2>&1); echo "${x}, ${d}"; done > v2_dates.csv
[osotolongo@detritus lfacehbi]$ join -t"," v0_dates.csv v2_dates.csv > long_dates.csv
[osotolongo@detritus lfacehbi]$ head long_dates.csv 
F001, 20141205, 20170124
F002, 20141205, 20170323
F003, 20141211, 20170123
F005, 20150107, 20170123
F006, 20141223, 20170124
F007, 20141219, 20170120
F008, 20141220, 20170125
F009, 20150110, 20170207
F010, 20150109, 20170208
F011, 20150110, 20170127

y para sacar la diferencia entre las fechas (es mas dificl de lo que parece), me escribo un script en perl,

get_years.pl

#!/usr/bin/perl
 
use strict;
use warnings;
use Date::Manip;
use Math::Round;
 
my $ifile = shift;
 
open IDF, "<$ifile" or die "No such file!";
while(<IDF>){
        my ($subject, $date1, $date2) = /^(.*), (\d+), (\d+)$/;
        my $d1 = ParseDate($date1);
        my $d2 = ParseDate($date2);
        my $diff = nearest(0.01, Delta_Format(DateCalc($d1,$d2),2,"%dh")/365.2425);
        print "$subject, $diff\n"; 
}

ya esta casi,

[osotolongo@detritus lfacehbi]$ ./get_years.pl long_dates.csv > facehbi_years_between.csv
[osotolongo@detritus lfacehbi]$ head facehbi_years_between.csv
F001, 2.14
F002, 2.3
F003, 2.12
F005, 2.05
F006, 2.09
F007, 2.09
F008, 2.1
F009, 2.08
F010, 2.08
F011, 2.05

A ver si pegamos esto,

[osotolongo@detritus lfacehbi]$ sed 's/,/;/' facehbi_years_between.csv > facehbi_years_between_sc.csv
[osotolongo@detritus lfacehbi]$ sed -i '1iPSubject; years' facehbi_years_between_sc.csv
[osotolongo@detritus lfacehbi]$ awk -F";" '{print $2";"$0}' flong.csv | sed 's/facehbi_//;s/SubjID_v0/Subject/' > flong_2concat.csv
[osotolongo@detritus lfacehbi]$ cp /nas/data/facehbi/facehbi_mri.csv ./
[osotolongo@detritus lfacehbi]$ sed -i '1iSubject;PSubject' facehbi_mri.csv
[osotolongo@detritus lfacehbi]$ join -t";" -1 2 -2 1 facehbi_mri.csv facehbi_years_between_sc.csv > fyears_2concat.csv
[osotolongo@detritus lfacehbi]$ sed 's/,/;/g;s/edat_v0/age/;s/Anyos_Escolaridad_FAC_v0/Education/;s/Sex_1H_0M_v0/Gender/' demographics.csv > dg_2concat.csv
[osotolongo@detritus lfacehbi]$ join -t";" fyears_2concat.csv dg_2concat.csv | sed 's/ //' > data_2concat.csv
 
[osotolongo@detritus lfacehbi]$ join -t";" -1 1 -2 2 flong_2concat.csv data_2concat.csv | tail -n +2 |awk -F";" '{print $3" "$2" 0 "$7" "$8" "$9"\n"$4" "$2" "$6" "$7+$6" "$8" "$9}' > qdec/long.qdec.table.dat 
[osotolongo@detritus lfacehbi]$ sed -i '1ifsid fsid-base years age education gender' qdec/long.qdec.table.dat
[osotolongo@detritus lfacehbi]$ head qdec/long.qdec.table.dat
fsid fsid-base years age education gender
facehbi_0001 flong_0001 0 71 8 0
f2cehbi_0001 flong_0001 2.14 73.14 8 0
facehbi_0002 flong_0002 0 70 12 1
f2cehbi_0002 flong_0002 2.3 72.3 12 1
facehbi_0003 flong_0003 0 70 8 0
f2cehbi_0003 flong_0003 2.12 72.12 8 0
facehbi_0005 flong_0005 0 68 20 1
f2cehbi_0004 flong_0005 2.05 70.05 20 1
facehbi_0006 flong_0006 0 64 14 0

y ya tenemos la tabla qdec lista

Two Stage Model

Antes de proceder aqui habria que comprobar que existen todos los sujetos y puntos de la tabla,

[osotolongo@detritus lfacehbi]$ for x in `tail -n +2 flong.csv | awk -F";" {'print $2".long."$1'}`; do ls -d /nas/data/subjects/${x}; done 
...
[osotolongo@detritus lfacehbi]$ for x in `tail -n +2 flong.csv | awk -F";" {'print $2'}`; do ls -d /nas/data/subjects/${x}; done
...
[osotolongo@detritus lfacehbi]$ for x in `tail -n +2 flong.csv | awk -F";" {'print $3".long."$1'}`; do ls -d /nas/data/subjects/${x}; done 
...
[osotolongo@detritus lfacehbi]$ for x in `tail -n +2 flong.csv | awk -F";" {'print $3'}`; do ls -d /nas/data/subjects/${x}; done
...

Si no hay ninguna linea de error en estos comandos, seguimos. En caso de faltar uno de estos directorios, hay que correr el analisis correspondiente de nuevo.

Ahora si,

[osotolongo@detritus lfacehbi]$ long_mris_slopes --sd /nas/data/subjects/ --qdec ./qdec/long.qdec.table.dat --meas thickness --hemi lh --do-avg --do-rate --do-pc1 --do-spc --do-stack --do-label --time years --qcache fsaverage

Esto es largo, dado el numero de sujetos, asi que mejor armarse de paciencia.

QDEC way

Ahora hay que preparar la tabla de QDEC,

$ long_qdec_table --qdec qdec/long.qdec.table.dat --cross --out qdec/cross.qdec.table.dat

y se ha de hacer un archov ~/.Qdec con la lista de variables,

MEASURE1 = long.thickness-avg
MEASURE2 = long.thickness-rate
MEASURE3 = long.thickness-pc1
MEASURE4 = long.thickness-spc

Ahora,

$ qdec --table qdec/cross.qdec.table.dat

y aqui, maniobrando con QDEC podemos obtener varios modelos.

GLM way

Aqui se miden los cambios en el grosor cortical ocurridos entre las dos visitas,

[osotolongo@detritus lfacehbi]$ long_mris_slopes --qdec qdec/long.qdec.table.dat --meas thickness --hemi lh --sd $SUBJECTS_DIR --do-pc1 --do-label --generic-time --fwhm 15 --qcache fsaverage --stack-pc1 lh.lfacehbi.thickness-pc1.stack.mgh --isec-labels lh.lfacehbi.fsaverage.cortex.label

Hacemos el modelo,

[osotolongo@detritus lfacehbi]$ mri_glmfit --osgm --glmdir lh.lfacehbi.thickness-pc1.fwhm15 --y lh.lfacehbi.thickness-pc1.stack.fwhm15.mgh --label lh.lfacehbi.fsaverage.cortex.label --surf fsaverage lh

Lo miramos,

freeview -f  /nas/data/subjects/fsaverage/surf/lh.inflated:annot=aparc.annot:annot_outline=1:overlay=lh.lfacehbi.thickness-pc1.fwhm15/osgm/sig.mgh:overlay_threshold=4,5 -viewport 3d

Corregimos por multiples comparaciones,

[osotolongo@detritus lfacehbi]$ mri_glmfit-sim --glmdir lh.lfacehbi.thickness-pc1.fwhm15/ --cache 4 neg --cwp  0.05 --2spaces

y vemos que clusters quedan,

freeview -f /nas/data/subjects/fsaverage/surf/lh.inflated:overlay=lh.lfacehbi.thickness-pc1.fwhm15/osgm/cache.th40.neg.sig.cluster.mgh:overlay_threshold=2,5:annot=lh.lfacehbi.thickness-pc1.fwhm15/osgm/cache.th40.neg.sig.ocn.annot -viewport 3d

La informacion de los clusters queda tambien almacenada,

[osotolongo@detritus lfacehbi]$ cat lh.lfacehbi.thickness-pc1.fwhm15/osgm/cache.th40.neg.sig.cluster.summary
# Cluster Growing Summary (mri_surfcluster)
# $Id: mri_surfcluster.c,v 1.57.2.3 2016/11/17 18:19:42 zkaufman Exp $
# $Id: mrisurf.c,v 1.781.2.6 2016/12/27 16:47:14 zkaufman Exp $
# CreationTime 2020/07/10-08:03:33-GMT
# cmdline mri_surfcluster.bin --in lh.lfacehbi.thickness-pc1.fwhm15//osgm/sig.mgh --mask lh.lfacehbi.thickness-pc1.fwhm15//mask.mgh --cwsig lh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.cluster.mgh --sum lh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.cluster.summary --ocn lh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.ocn.mgh --annot aparc --cwpvalthresh 0.05 --o lh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.masked.mgh --no-fixmni --csd /usr/local/freesurfer/average/mult-comp-cor/fsaverage/lh/cortex/fwhm17/neg/th40/mc-z.csd --csdpdf lh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.pdf.dat --vwsig lh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.voxel.mgh --vwsigmax lh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.voxel.max.dat --oannot lh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.ocn.annot --bonferroni 2 --surf white 
# cwd /nas/data/lfacehbi
# sysname  Linux
# hostname detritus.fundacioace.com
# machine  x86_64
# FixVertexAreaFlag 1
# FixSurfClusterArea 1
# 
# Input      lh.lfacehbi.thickness-pc1.fwhm15//osgm/sig.mgh
# Frame Number      0
# srcsubj fsaverage
# hemi lh
# surface white
# group_avg_surface_area 82220
# group_avg_vtxarea_loaded 1
# annot aparc
# SUBJECTS_DIR /nas/data/subjects
# SearchSpace_mm2 75384.9
# SearchSpace_vtx 147377
# Bonferroni 2
# Minimum Threshold 4
# Maximum Threshold infinity
# Threshold Sign    neg
# AdjustThreshWhenOneTail 1
# CW PValue Threshold: 0.05 
# Area Threshold    0 mm^2
# CSD thresh  4.000000
# CSD nreps    10000
# CSD simtype  null-z
# CSD contrast NA
# CSD confint  90.000000
# Overall max 3.84849 at vertex 122407
# Overall min -6.15084 at vertex 106728
# NClusters          6
# FixMNI = 0
# 
# ClusterNo  Max   VtxMax   Size(mm^2)  MNIX   MNIY   MNIZ    CWP    CWPLow    CWPHi   NVtxs    WghtVtx   Annot
   1       -5.335  111613    886.69    -44.2  -69.7    9.3  0.00020  0.00000  0.00040   1561    -6564.19  inferiorparietal
   2       -6.151  106728    721.02    -43.3  -56.5   25.3  0.00020  0.00000  0.00040   1591    -7085.05  inferiorparietal
   3       -5.789  104223    577.10    -43.7  -39.2   11.4  0.00020  0.00000  0.00040   1440    -6316.08  superiortemporal
   4       -5.241  148689    225.71    -57.9  -33.7  -16.5  0.00080  0.00040  0.00140    355    -1496.33  middletemporal
   5       -4.453  135429    105.36    -51.6  -34.7    0.1  0.02938  0.02642  0.03253    249     -997.82  bankssts
   6       -4.944   72121    102.69    -40.1  -31.3   35.2  0.03155  0.02840  0.03469    321    -1362.45  supramarginal

Para el hemisferio derecho se hace exactamente los mismo

[osotolongo@detritus lfacehbi]$ cat rh.lfacehbi.thickness-pc1.fwhm15/osgm/cache.th40.neg.sig.cluster.summary 
# Cluster Growing Summary (mri_surfcluster)
# $Id: mri_surfcluster.c,v 1.57.2.3 2016/11/17 18:19:42 zkaufman Exp $
# $Id: mrisurf.c,v 1.781.2.6 2016/12/27 16:47:14 zkaufman Exp $
# CreationTime 2020/07/11-10:21:49-GMT
# cmdline mri_surfcluster.bin --in rh.lfacehbi.thickness-pc1.fwhm15//osgm/sig.mgh --mask rh.lfacehbi.thickness-pc1.fwhm15//mask.mgh --cwsig rh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.cluster.mgh --sum rh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.cluster.summary --ocn rh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.ocn.mgh --annot aparc --cwpvalthresh 0.05 --o rh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.masked.mgh --no-fixmni --csd /usr/local/freesurfer/average/mult-comp-cor/fsaverage/rh/cortex/fwhm17/neg/th40/mc-z.csd --csdpdf rh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.pdf.dat --vwsig rh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.voxel.mgh --vwsigmax rh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.voxel.max.dat --oannot rh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.ocn.annot --bonferroni 2 --surf white 
# cwd /nas/data/lfacehbi
# sysname  Linux
# hostname detritus.fundacioace.com
# machine  x86_64
# FixVertexAreaFlag 1
# FixSurfClusterArea 1
# 
# Input      rh.lfacehbi.thickness-pc1.fwhm15//osgm/sig.mgh
# Frame Number      0
# srcsubj fsaverage
# hemi rh
# surface white
# group_avg_surface_area 82167.6
# group_avg_vtxarea_loaded 1
# annot aparc
# SUBJECTS_DIR /nas/data/subjects
# SearchSpace_mm2 74448.5
# SearchSpace_vtx 146051
# Bonferroni 2
# Minimum Threshold 4
# Maximum Threshold infinity
# Threshold Sign    neg
# AdjustThreshWhenOneTail 1
# CW PValue Threshold: 0.05 
# Area Threshold    0 mm^2
# CSD thresh  4.000000
# CSD nreps    10000
# CSD simtype  null-z
# CSD contrast NA
# CSD confint  90.000000
# Overall max 2.48051 at vertex 151862
# Overall min -10.6092 at vertex 53693
# NClusters          8
# FixMNI = 0
# 
# ClusterNo  Max   VtxMax   Size(mm^2)  MNIX   MNIY   MNIZ    CWP    CWPLow    CWPHi   NVtxs    WghtVtx   Annot
   1      -10.609   53693   3009.16     51.4    1.9  -28.1  0.00020  0.00000  0.00040   5686   -29777.36  middletemporal
   2       -5.368   66629    232.21     19.8   37.8   40.7  0.00060  0.00020  0.00100    352    -1559.05  superiorfrontal
   3       -4.418    3639    191.12     50.5    3.8   17.7  0.00260  0.00180  0.00360    415    -1653.58  precentral
   4       -4.710  145684    151.25     44.1   29.2   12.8  0.00878  0.00719  0.01057    275    -1142.94  rostralmiddlefrontal
   5       -6.086  138124    149.07     25.4  -11.9   48.9  0.00898  0.00739  0.01077    384    -1799.79  precentral
   6       -4.892   48409    137.53     22.1   12.8   48.0  0.01296  0.01097  0.01494    304    -1282.61  superiorfrontal
   7       -5.038  154726     98.25      9.8   62.1    2.0  0.03115  0.02800  0.03430    123     -527.80  superiorfrontal
   8       -4.289  152608     83.01     36.8    3.4   35.4  0.04898  0.04508  0.05288    147     -570.54  caudalmiddlefrontal

añadiendo variables

[osotolongo@detritus lfacehbi]$ cp /nas/data/facehbi/fbb_results/fbb_centiloid.csv fbb_centiloid_v0.csv
[osotolongo@detritus lfacehbi]$ cp /nas/data/f2cehbi/fbb_results/fbb_centiloid.csv fbb_centiloid_v2.csv
[osotolongo@detritus lfacehbi]$ awk -F "," '{print $1";"$3}' fbb_centiloid_v0.csv | sed 's/Centiloid/CLv0/' > centiloid_v0.csv
[osotolongo@detritus lfacehbi]$ awk -F "," '{print $1";"$3}' fbb_centiloid_v2.csv | sed 's/Centiloid/CLv2/' > centiloid_v2.csv
[osotolongo@detritus lfacehbi]$ head data_2concat.csv 
PSubject;Subject;years;age;Education;Gender
F001;0001;2.14;71;8;0
F002;0002;2.3;70;12;1
F003;0003;2.12;70;8;0
F005;0005;2.05;68;20;1
F006;0006;2.09;64;14;0
F007;0007;2.09;59;19;1
F008;0008;2.1;55;16;0
F009;0009;2.08;67;16;0
F010;0010;2.08;68;20;1
[osotolongo@detritus lfacehbi]$ head centiloid_v0.csv
Subject; CLv0
0001;-5.5932502480232
0002;12.0568081099059
0003;-7.34233394547346
0004;10.1249519364766
0005;-8.16586361015402
0006;37.3631081538974
0007;75.8928076873738
0008;-17.1955789921458
0009;3.61335051283359
[osotolongo@detritus lfacehbi]$ join -t";" -1 2 -2 1 data_2concat.csv centiloid_v0.csv > data_2concat_temp.csv
[osotolongo@detritus lfacehbi]$ head data_2concat_temp.csv 
Subject;PSubject;years;age;Education;Gender; CLv0
0001;F001;2.14;71;8;0;-5.5932502480232
0002;F002;2.3;70;12;1;12.0568081099059
0003;F003;2.12;70;8;0;-7.34233394547346
0005;F005;2.05;68;20;1;-8.16586361015402
0006;F006;2.09;64;14;0;37.3631081538974
0007;F007;2.09;59;19;1;75.8928076873738
0008;F008;2.1;55;16;0;-17.1955789921458
0009;F009;2.08;67;16;0;3.61335051283359
0010;F010;2.08;68;20;1;-11.4216448690239
[osotolongo@detritus lfacehbi]$ join -t";" data_2concat_temp.csv centiloid_v2.csv  > data_2concat_extra.csv
[osotolongo@detritus lfacehbi]$ head data_2concat_extra.csv 
Subject;PSubject;years;age;Education;Gender; CLv0; CLv2
0001;F001;2.14;71;8;0;-5.5932502480232;-3.01177159588474
0002;F002;2.3;70;12;1;12.0568081099059;9.07793600565404
0003;F003;2.12;70;8;0;-7.34233394547346;-7.74822070800795
0005;F005;2.05;68;20;1;-8.16586361015402;39.4041710834628
0006;F006;2.09;64;14;0;37.3631081538974;80.3079451944081
0007;F007;2.09;59;19;1;75.8928076873738;-20.4246399521569
0008;F008;2.1;55;16;0;-17.1955789921458;5.61946211270271
0009;F009;2.08;67;16;0;3.61335051283359;-14.8767332876036
0010;F010;2.08;68;20;1;-11.4216448690239;0.505619291014511
[osotolongo@detritus lfacehbi]$ head facehbi_apoe4_sorted.csv 
PSubject;ApoeE4
F001;1
F002;0
F003;0
F004;0
F005;0
F006;1
F007;0
F008;0
F009;0
[osotolongo@detritus lfacehbi]$ join -t";" -1 2 -2 1 data_2concat_extra.csv facehbi_apoe4_sorted.csv > data_2concat_wvars.csv
[osotolongo@detritus lfacehbi]$ head data_2concat_wvars.csv
PSubject;Subject;years;age;Education;Gender; CLv0; CLv2;ApoeE4
F001;0001;2.14;71;8;0;-5.5932502480232;-3.01177159588474;1
F002;0002;2.3;70;12;1;12.0568081099059;9.07793600565404;0
F003;0003;2.12;70;8;0;-7.34233394547346;-7.74822070800795;0
F005;0005;2.05;68;20;1;-8.16586361015402;39.4041710834628;0
F006;0006;2.09;64;14;0;37.3631081538974;80.3079451944081;1
F007;0007;2.09;59;19;1;75.8928076873738;-20.4246399521569;0
F008;0008;2.1;55;16;0;-17.1955789921458;5.61946211270271;0
F009;0009;2.08;67;16;0;3.61335051283359;-14.8767332876036;0
[osotolongo@detritus lfacehbi]$ head flong_2concat.csv 
Subject;flong_subject;SubjID_v0;SubjID_v2
0001;flong_0001;facehbi_0001;f2cehbi_0001
0002;flong_0002;facehbi_0002;f2cehbi_0002
0003;flong_0003;facehbi_0003;f2cehbi_0003
0005;flong_0005;facehbi_0005;f2cehbi_0004
0006;flong_0006;facehbi_0006;f2cehbi_0005
0007;flong_0007;facehbi_0007;f2cehbi_0006
0008;flong_0008;facehbi_0008;f2cehbi_0007
0009;flong_0009;facehbi_0009;f2cehbi_0008
0010;flong_0010;facehbi_0010;f2cehbi_0009
[osotolongo@detritus lfacehbi]$ join -t";" -1 1 -2 2 flong_2concat.csv data_2concat_wvars.csv | tail -n +2 |awk -F";" '{print $3" "$2" 0 "$7" "$8" "$9" "$10" "$12"\n"$4" "$2" "$6" "$7+$6" "$8" "$9" "$11" "$12}'  > qdec/long.qdec.table.extra.dat
[osotolongo@detritus lfacehbi]$ sed -i '1ifsid fsid-base years age education gender centiloid apoe' qdec/long.qdec.table.extra.dat
[osotolongo@detritus lfacehbi]$ head qdec/long.qdec.table.extra.dat
fsid fsid-base years age education gender centiloid apoe
facehbi_0001 flong_0001 0 71 8 0 -5.5932502480232 1
f2cehbi_0001 flong_0001 2.14 73.14 8 0 -3.01177159588474 1
facehbi_0002 flong_0002 0 70 12 1 12.0568081099059 0
f2cehbi_0002 flong_0002 2.3 72.3 12 1 9.07793600565404 0
facehbi_0003 flong_0003 0 70 8 0 -7.34233394547346 0
f2cehbi_0003 flong_0003 2.12 72.12 8 0 -7.74822070800795 0
facehbi_0005 flong_0005 0 68 20 1 -8.16586361015402 0
f2cehbi_0004 flong_0005 2.05 70.05 20 1 39.4041710834628 0
facehbi_0006 flong_0006 0 64 14 0 37.3631081538974 1

Voy a hacer el run de nuevo par asegurarme de que sean los mismos sujetos. ¿Porqué? Porque si falta algun valor de Centiloid o de APOE la tabla es distinta de la plantilla de sujetos y el analisis no sera valido.

[osotolongo@detritus lfacehbi]$ long_mris_slopes --qdec qdec/long.qdec.table.extra.dat --meas thickness --hemi lh --sd 
$SUBJECTS_DIR --do-pc1 --do-label --generic-time --fwhm 15 --qcache fsaverage --stack-pc1 lh.lfacehbi.thickness-pc1.sta
ck.mgh --isec-labels lh.lfacehbi.fsaverage.cortex.label
...
[osotolongo@detritus lfacehbi]$ mri_glmfit --osgm --glmdir lh.lfacehbi.thickness-pc1.fwhm15 --y lh.lfacehbi.thickness-pc1.stack.fwhm15.mgh --label lh.lfacehbi.fsaverage.cortex.label --surf fsaverage lh
[osotolongo@detritus lfacehbi]$ mri_glmfit-sim --glmdir lh.lfacehbi.thickness-pc1.fwhm15/ --cache 4 neg --cwp  0.05 --2spaces

[osotolongo@detritus lfacehbi]$ cat lh.lfacehbi.thickness-pc1.fwhm15/osgm/cache.th40.neg.sig.cluster.summary
# Cluster Growing Summary (mri_surfcluster)
# $Id: mri_surfcluster.c,v 1.57.2.3 2016/11/17 18:19:42 zkaufman Exp $
# $Id: mrisurf.c,v 1.781.2.6 2016/12/27 16:47:14 zkaufman Exp $
# CreationTime 2020/07/16-07:40:57-GMT
# cmdline mri_surfcluster.bin --in lh.lfacehbi.thickness-pc1.fwhm15//osgm/sig.mgh --mask lh.lfacehbi.thickness-pc1.fwhm15//mask.mgh --cwsig lh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.cluster.mgh --sum lh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.cluster.summary --ocn lh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.ocn.mgh --annot aparc --cwpvalthresh 0.05 --o lh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.masked.mgh --no-fixmni --csd /usr/local/freesurfer/average/mult-comp-cor/fsaverage/lh/cortex/fwhm16/neg/th40/mc-z.csd --csdpdf lh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.pdf.dat --vwsig lh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.voxel.mgh --vwsigmax lh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.voxel.max.dat --oannot lh.lfacehbi.thickness-pc1.fwhm15//osgm/cache.th40.neg.sig.ocn.annot --bonferroni 2 --surf white 
# cwd /nas/data/lfacehbi
# sysname  Linux
# hostname detritus.fundacioace.com
# machine  x86_64
# FixVertexAreaFlag 1
# FixSurfClusterArea 1
# 
# Input      lh.lfacehbi.thickness-pc1.fwhm15//osgm/sig.mgh
# Frame Number      0
# srcsubj fsaverage
# hemi lh
# surface white
# group_avg_surface_area 82220
# group_avg_vtxarea_loaded 1
# annot aparc
# SUBJECTS_DIR /nas/data/subjects
# SearchSpace_mm2 75466.5
# SearchSpace_vtx 147613
# Bonferroni 2
# Minimum Threshold 4
# Maximum Threshold infinity
# Threshold Sign    neg
# AdjustThreshWhenOneTail 1
# CW PValue Threshold: 0.05 
# Area Threshold    0 mm^2
# CSD thresh  4.000000
# CSD nreps    10000
# CSD simtype  null-z
# CSD contrast NA
# CSD confint  90.000000
# Overall max 3.29752 at vertex 122410
# Overall min -6.1539 at vertex 79545
# NClusters          5
# FixMNI = 0
# 
# ClusterNo  Max   VtxMax   Size(mm^2)  MNIX   MNIY   MNIZ    CWP    CWPLow    CWPHi   NVtxs    WghtVtx   Annot
   1       -4.666   69976    543.01    -46.8  -63.9   -5.4  0.00020  0.00000  0.00040    950    -3790.21  inferiortemporal
   2       -6.154   79545    343.66    -59.2  -33.7  -16.6  0.00020  0.00000  0.00040    554    -2475.20  middletemporal
   3       -5.453  122049    309.70    -43.5  -57.4   26.6  0.00020  0.00000  0.00040    664    -2881.33  inferiorparietal
   4       -5.080  120104    241.40    -42.6  -38.3   10.1  0.00040  0.00000  0.00080    683    -2914.49  superiortemporal
   5       -5.054   72121    120.14    -40.1  -31.3   35.2  0.01475  0.01256  0.01693    367    -1580.34  supramarginal

Voy a intentar algo aqui

[osotolongo@detritus lfacehbi]$ sed 's/fsid fsid-base/Variables/;s/facehbi_\([0-9]*\) flong_\([0-9]*\)/Input facehbi_\1.long.flong_\2 Main/;s/f2cehbi_\([0-9]*\) flong_\([0-9]*\)/Input f2cehbi_\1.long.flong_\2 Main/' qdec/long.qdec.table.extra.dat > qdec.body    
[osotolongo@detritus lfacehbi]$ cat headers.txt qdec.body > qdec.fsgd
                                                                                                           [osotolongo@detritus lfacehbi]$ head qdec.fsgd
Variables years age education gender centiloid apoe
Input facehbi_0001.long.flong_0001 Main 0 71 8 0 -5.5932502480232 1
Input f2cehbi_0001.long.flong_0001 Main 2.14 73.14 8 0 -3.01177159588474 1
Input facehbi_0002.long.flong_0002 Main 0 70 12 1 12.0568081099059 0
Input f2cehbi_0002.long.flong_0002 Main 2.3 72.3 12 1 9.07793600565404 0
Input facehbi_0003.long.flong_0003 Main 0 70 8 0 -7.34233394547346 0
Input f2cehbi_0003.long.flong_0003 Main 2.12 72.12 8 0 -7.74822070800795 0
Input facehbi_0005.long.flong_0005 Main 0 68 20 1 -8.16586361015402 0
Input f2cehbi_0004.long.flong_0005 Main 2.05 70.05 20 1 39.4041710834628 0
Input facehbi_0006.long.flong_0006 Main 0 64 14 0 37.3631081538974 1
 
[osotolongo@detritus lfacehbi]$ cat age.mtx
0 0 1 0 0 0 0
[osotolongo@detritus lfacehbi]$ cat apoe.mtx 
0 0 0 0 0 0 1
[osotolongo@detritus lfacehbi]$ cat centiloid.mtx
0 0 0 0 0 1 0
[osotolongo@detritus lfacehbi]$ cat years.mtx
0 1 0 0 0 0 0

Primero voy a hacer el cache,

[osotolongo@detritus lfacehbi]$ mris_preproc --fsgd qdec.fsgd --target fsaverage --hemi lh --meas thickness --out lh.long.thickness.00.mgh
[osotolongo@detritus lfacehbi]$ mri_surf2surf --hemi lh --s fsaverage --sval lh.long.thickness.00.mgh --fwhm 10 --cortex --tval lh.long.thickness.10.mgh

y lanzo el analisis ahora,

[osotolongo@detritus lfacehbi]$ mri_glmfit --fsgd qdec.fsgd --glmdir lh.long.thickness.fwhm10 --y lh.long.thickness.10.mgh --C centiloid.mtx --C apoe.mtx --C years.mtx --C age.mtx --surf fsaverage lh

Voy a ver,

Edad:

$ freeview -f  /nas/data/subjects/fsaverage/surf/lh.inflated:annot=aparc.annot:annot_outline=1:overlay=lh.long.thickness.fwhm10/age/sig.mgh:overlay_threshold=4,5 -viewport 3d

APOEe4:

$ freeview -f  /nas/data/subjects/fsaverage/surf/lh.inflated:annot=aparc.annot:annot_outline=1:overlay=lh.long.thickness.fwhm10/apoe/sig.mgh:overlay_threshold=4,5 -viewport 3d

Paired Analysis

Esto es una especie de ANOVA simplificada para dos puntos.

https://surfer.nmr.mgh.harvard.edu/fswiki/PairedAnalysis

Variables:

[osotolongo@detritus lfacehbi]$ join -t";" -1 1 -2 2 flong_2concat.csv data_2concat_wvars.csv | head -n 1 | sed 's/;/\n/g' | cat -n
     1	Subject
     2	flong_subject
     3	SubjID_v0
     4	SubjID_v2
     5	PSubject
     6	years
     7	age
     8	Education
     9	Gender
    10	 CLv0
    11	 CLv2
    12	ApoeE4

Los pares son asi entonces,

[osotolongo@detritus lfacehbi]$ join -t";" -1 1 -2 2 flong_2concat.csv data_2concat_wvars.csv | tail -n +2 |awk -F";" '{print "Input "$3" Main "$7" "$8" "$9" "$10" "$12"\nInput "$4" Main "$7+$6" "$8" "$9" "$11" "$12}' > pairs.fsgd
[osotolongo@detritus lfacehbi]$ sed -i '1iGroupDescriptorFile 1\nClass Main' pairs.fsgd
[osotolongo@detritus lfacehbi]$ head pairs.fsgd
GroupDescriptorFile 1
Class Main
Input facehbi_0001 Main 71 8 0 -5.5932502480232 1
Input f2cehbi_0001 Main 73.14 8 0 -3.01177159588474 1
Input facehbi_0002 Main 70 12 1 12.0568081099059 0
Input f2cehbi_0002 Main 72.3 12 1 9.07793600565404 0
Input facehbi_0003 Main 70 8 0 -7.34233394547346 0
Input f2cehbi_0003 Main 72.12 8 0 -7.74822070800795 0
Input facehbi_0005 Main 68 20 1 -8.16586361015402 0
Input f2cehbi_0004 Main 70.05 20 1 39.4041710834628 0

y empezamos,

[osotolongo@detritus lfacehbi]$ mris_preproc --target fsaverage --hemi lh --meas thickness --out lh.paired-diff.thickness.mgh --fsgd pairs.fsgd --paired-diff
nsubjects = 318
tmpdir is ./tmp.mris_preproc.89542
/nas/data/lfacehbi
Log file is lh.paired-diff.thickness.mris_preproc.log
...
...
...
[osotolongo@detritus lfacehbi]$ mri_surf2surf --s fsaverage --hemi lh --fwhm 5 --sval lh.paired-diff.thickness.mgh --tval lh.paired-diff.thickness.sm05.mgh
...
...
...

hacemos los .mtx,

[osotolongo@detritus lfacehbi]$ echo "0 1 0 0 0 0" > age.mtx

cada variable a analizar debe tener uno,

[osotolongo@detritus lfacehbi]$ cat mean.mtx
1 0 0 0 0 0
[osotolongo@detritus lfacehbi]$ cat age.mtx
0 1 0 0 0 0
[osotolongo@detritus lfacehbi]$ cat apoe.mtx
0 0 0 0 0 1
[osotolongo@detritus lfacehbi]$ cat centiloid.mtx
0 0 0 0 1 0

y hay que hacer el segundo archivo .fsgd. Primero voy a eliminar lso caracteres raros que de alguan manera han venido del archivo windows original.

[osotolongo@detritus lfacehbi]$ sed 's/^M//' pairs.fsgd > pairs_unix.fsgd
[osotolongo@detritus lfacehbi]$ tr -d '\r' < pairs_unix.fsgd > pairs_ok.fsgd

y ahora elimino las referencias al segundo punto y cambio el nombre del primer punto, para tener nombres que no existan aun,

[osotolongo@detritus lfacehbi]$ grep -v  f2cehbi pairs_ok.fsgd | sed 's/facehbi/facehbi_pair/' > pair_diff.fsgd
[osotolongo@detritus lfacehbi]$ sed -i '3iVariables Age Education Gender CL APOE' pair_diff.fsgd

y ahora el glmfit,

[osotolongo@detritus lfacehbi]$ mri_glmfit --glmdir lh.paired-diff --y lh.paired-diff.thickness.sm05.mgh --fsgd pair_diff.fsgd --C mean.mtx --C age.mtx --C apoe.mtx --C centiloid.mtx --surf fsaverage lh

y obtenemos un analisis para cada variable,

[osotolongo@detritus lfacehbi]$ tree -d lh.paired-diff/
lh.paired-diff/
├── age
├── apoe
├── centiloid
└── mean
 
4 directories
[osotolongo@detritus lfacehbi]$ tree lh.paired-diff/mean/
lh.paired-diff/mean/
├── C.dat
├── cnr.mgh
├── efficiency.dat
├── F.mgh
├── gamma.mgh
├── gammavar.mgh
├── maxvox.dat
├── pcc.mgh
├── sig.mgh
└── z.mgh
 
0 directories, 10 files

Desgraciadamente sin resultados

Detritus Wiki

Table of Contents