Correlational Tractography with Large Number of Subjects (Nothing Output)

[jamiela...@gmail.com]

Hello all,

I was trying to complete whole-brain correlational tractography with a large sample (N=707) and kept running into errors and DSI-studio crashes. I've tried via the GUI and command line (both singularity), but things always crash/fail. I tried it on a Linux machine (Ubuntu 16), as well as my MacBook Pro (running BigSur). I searched in the archive of the group, but nothing similar popped up (though perhaps I missed things?). 

In a recent iteration using the command line, I used this command and flags--

singularity exec dsistudio_latest.sif dsi_studio --action=cnt --source=connectometry.qa.db.fib.gz --demo=HBN_APQ.csv --variable_list=0 --voi=0 --permutation=50 --fdr_threshold=2.5 --length_threshold=20 --t_threshold=2.5 --normalized_qa=1 --tip=4

(* I know that permutations were very low... I was trying to see if things run, etc.). 

The command line moves forward, but then nothing is output; here's my command line output--

DSI Studio Jan 23 2021, Fang-Cheng Yeh

QStandardPaths: XDG_RUNTIME_DIR points to non-existing path '/run/user/1002', please create it with 0700 permissions.

Starting GUI-based command line interface.

source=connectometry.qa.db.fib.gz

action=cnt

reading connectometry db

process 183406: D-Bus library appears to be incorrectly set up; failed to read machine uuid: UUID file '/etc/machine-id' should contain a hex string of length 32, not length 0, with no other text

See the manual page for dbus-uuidgen to correct this issue.

libGL error: No matching fbConfigs or visuals found

libGL error: failed to load driver: swrast

demo=HBN_APQ.csv

variable_list=0

voi=0

variables=APQ_P_ID

study variable=APQ_P_ID

t_threshold=2.5

tip=4

nonparametric=1

permutation=50

length_threshold=20

select=

normalized_qa=1

fdr_threshold=2.5

running connectometry

output results

. A nonparametric Spearman correlation was used to derive the correlation. A total of 707 subjects were included in the analysis. A T-score threshold of 2.5 was assigned and tracked using a deterministic fiber tracking algorithm (Yeh et al. PLoS ONE 8(11): e80713, 2013) to obtain correlational tractography. The QA values were normalized. The tracks were filtered by topology-informed pruning (Yeh et al. Neurotherapeutics, 16(1), 52-58, 2019) with 4 iteration(s). An FDR threshold of 0.5 was used to select tracks. To estimate the false discovery rate, a total of 100 randomized permutations were applied to the group label to obtain the null distribution of the track length.

loading /opt/dsi-studio/dsi_studio_64/track/HCP1065.tt.gz

warping atlas tracks to subject space

Unrecognized OpenGL version

Unrecognized OpenGL version

QOpenGLTexture::setData() requires a valid current context

**** I wasn't sure if my command line syntax was incorrect, etc., or if there was something else incorrect in what I was doing. Thoughts? 

I uploaded my connectometry database and a csv file to the DSI-studio DropBox (as the folder Hanson_DSIstudio, with both files). I'm figuring my command line syntax was incorrect, or my linux system had some dependency issues? Given that I was trying to run command line analyses, I didn't think the OpenGL (graphics) issues should matter? But I thought I would follow-up to see if others have run into this, or had thoughts on what I might be doing wrong. 

Many thanks for the help, 

Jamie.

[Frank Yeh]

I am sorry about this problem and will check out the data and fix it asap.

The command line for correlational tractography will need OpenGL to generate tractogram report.

It is likely to be the cause that it fails on Linux.

Frank

--
You received this message because you are subscribed to the Google Groups "DSI Studio" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dsi-studio+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/1b709ff7-2c97-4c29-9ff5-13488c12c349o%40googlegroups.com.

[Jamie Hanson]

Hello all--

I wanted to post more about this to see if others had run into (and/or worked around this type of issue). Running things locally on my mac is fine, but takes a good bit of time (~18-20 hours depending on the number of permutations for whole brain analyses). So, I was trying to leverage the Dell PowerEdge R830 server my group has, it is running Ubuntu 16 and has a fair number of cores (so that is useful for permuting and this reasonable number of subjects). Running the GUI as singularity image crashes on my server. That error/output is something along the lines of--

****

singularity exec dsistudio_latest.sif dsi_studio

QStandardPaths: XDG_RUNTIME_DIR points to non-existing path '/run/user/1002', please create it with 0700 permissions.

process 166014: D-Bus library appears to be incorrectly set up; failed to read machine uuid: UUID file '/etc/machine-id' should contain a hex string of length 32, not length 0, with no other text

See the manual page for dbus-uuidgen to correct this issue.
libGL error: No matching fbConfigs or visuals found
libGL error: failed to load driver: swrast

reading connectometry db
****
(Once I try to load the covariate list, DSI-studio crashes)

If I use the command line, things look to complete with ~80 cores being used for about ~2 hours. I run the following syntax--
singularity exec /home/jamielh/Volumes/Hanson/Public_Access_MRI/NKI_HealthyBrainNetwork/Neuroimaging/BIDS/Derived/dsi-studio/dsistudio_latest.sif dsi_studio --action=cnt --source=connectometry.qa.db.fib.gz --demo=FSQ_combined.csv --variable_list=0,1,2,3,4,5,6 --voi=6 --permutation=8000 --fdr_threshold=2.5 --length_threshold=20 --t_threshold=2.5 --normalized_qa=1 --tip=4 --exclude_cb=1

And get the following output--
****
DSI Studio Feb  4 2021, Fang-Cheng Yeh

QStandardPaths: XDG_RUNTIME_DIR points to non-existing path '/run/user/1002', please create it with 0700 permissions.
Starting GUI-based command line interface.
source=connectometry.qa.db.fib.gz
action=cnt
reading connectometry db

process 156512: D-Bus library appears to be incorrectly set up; failed to read machine uuid: UUID file '/etc/machine-id' should contain a hex string of length 32, not length 0, with no other text

See the manual page for dbus-uuidgen to correct this issue.
libGL error: No matching fbConfigs or visuals found
libGL error: failed to load driver: swrast

demo=FSQ_combined.csv
variable_list=0,1,2,3,4,5,6
voi=6
variables=Agemc,AgemcSQ,Sex,CBICsitecode,RUsitecode,SIsitecode,FSQ04r
study variable=FSQ04r
t_threshold=2.5
tip=4
nonparametric=1
permutation=8000
length_threshold=20
select=
normalize_qa=1
exclude_cb=1
fdr_threshold=2.5
running connectometry
read region from FreeSurferSeg:Left-Cerebellum-White-Matter
searching FreeSurferSeg from the atlas pool...
loading atlas FreeSurferSeg from the atlas list.
loading Left-Cerebellum-White-Matter from FreeSurferSeg atlas
read region from FreeSurferSeg:Left-Cerebellum-Cortex
searching FreeSurferSeg from the atlas pool...
loading atlas FreeSurferSeg from the atlas list.
loading Left-Cerebellum-Cortex from FreeSurferSeg atlas
read region from FreeSurferSeg:Right-Cerebellum-White-Matter
searching FreeSurferSeg from the atlas pool...
loading atlas FreeSurferSeg from the atlas list.
loading Right-Cerebellum-White-Matter from FreeSurferSeg atlas
read region from FreeSurferSeg:Right-Cerebellum-Cortex
searching FreeSurferSeg from the atlas pool...
loading atlas FreeSurferSeg from the atlas list.
loading Right-Cerebellum-Cortex from FreeSurferSeg atlas
.t2.nqa.fdr0.500000..cerebellum.pos_corr.tt.gz
.t2.nqa.fdr0.500000..cerebellum.neg_corr.tt.gz
output results

. A nonparametric Spearman partial correlation was used to derive the correlation, and the effect of Agemc, AgemcSQ, Sex, CBICsitecode, RUsitecode, and SIsitecode was removed using a multiple regression model. A total of 413 subjects were included in the analysis. A T-score threshold of 2.5 was assigned and tracked using a deterministic fiber tracking algorithm (Yeh et al. PLoS ONE 8(11): e80713, 2013) to obtain correlational tractography. The QA values were normalized. The tracks were filtered by topology-informed pruning (Yeh et al. Neurotherapeutics, 16(1), 52-58, 2019) with 4 iteration(s). An FDR threshold of 0.5 was used to select tracks. To estimate the false discovery rate, a total of 8000 randomized permutations were applied to the group label to obtain the null distribution of the track length. The tracking algorithm used cerebellum as the terminating region.

loading /opt/dsi-studio/dsi_studio_64/track/HCP1065.tt.gz
warping atlas tracks to subject space
Unrecognized OpenGL version
Unrecognized OpenGL version
QOpenGLTexture::setData() requires a valid current context
****

At that point, things return back to an open/active command prompt, but no files/results are output. I looked in /run/user/1002 (which actually does exist), but there's no result files there), etc. I also wondered about the output ".t2.nqa.fdr0.500000..cerebellum.pos_corr.tt.gz" and it potentially being a hidden file (but that wasn't true). I'm figuring this is all OpenGL related. I keep diving into the host of pages about troubleshooting OpenGL errors, mainly trying to remove some mesa and NVIDIA drivers (but nothing has been successful yet). Other simple things that I should try to troubleshoot, etc.?

All the best,
Jamie.

You received this message because you are subscribed to a topic in the Google Groups "DSI Studio" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/dsi-studio/9i80QbhZa2k/unsubscribe.
To unsubscribe from this group and all its topics, send an email to dsi-studio+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/CAG_QrtKV8h%2Bzv%2BvN3v9cnU495B3W1gppW6Yx73t-oOfSijitfQ%40mail.gmail.com.

[Frank Yeh]

I updated DSI Studio and now you can add --no_tractogram=1 to skip the
3D rendering of the tracks.
Once the analysis is done, DSI Studio will still output tractography
file, and you can use Mac or Windows to create tractograms separately.
Hope this works for you.
Frank

> To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/CAN8Mcn8_GsuQrDeWA0gPXDPwgZ1dWZWOfTo9tcZWrMX_8hH6vQ%40mail.gmail.com.

[Jamie Hanson]

Oh, yes! I think this will be very helpful. 

I can work independently/after analyses to look at the results.

Wow, many thanks for all this help! It is really invaluable!

To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/CAG_QrtK5RXM5XRpwftS7ufZv%2BTme8NoXD_gJCaOLvZWWqrjzbg%40mail.gmail.com.

[Frank Yeh]

Its my pleasure to help!
Frank

> To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/CAN8Mcn-te8U19YmqJD%3D8C7cfDedBSLQr%2BAXsMCyKmq0hKfBWrA%40mail.gmail.com.

[Jamie Hanson]

Still running into issues, I used the syntax (after pulling a new singularity imag)--

singularity exec dsistudio_latest.sif dsi_studio --action=cnt --source=connectometry.qa.db.fib.gz --demo=APQ_combined.csv --variable_list=0,1,2,3,4,5,6 --voi=6 --permutation=4000 --fdr_threshold=2.5 --length_threshold=20 --t_threshold=2.5 --normalized_qa=1 --tip=4 --exclude_cb=1 --no_tractogram=1

*** 

and got the following output--

*** 

DSI Studio Feb  5 2021, Fang-Cheng Yeh

QStandardPaths: XDG_RUNTIME_DIR points to non-existing path '/run/user/1002', please create it with 0700 permissions.

Starting GUI-based command line interface.

source=connectometry.qa.db.fib.gz

action=cnt

reading connectometry db

process 175221: D-Bus library appears to be incorrectly set up; failed to read machine uuid: UUID file '/etc/machine-id' should contain a hex string of length 32, not length 0, with no other text

See the manual page for dbus-uuidgen to correct this issue.

libGL error: No matching fbConfigs or visuals found

libGL error: failed to load driver: swrast

demo=APQ_combined.csv

variable_list=0,1,2,3,4,5,6

voi=6

variables=Age,Age_Sq,Sex,dummyCode_CBIC,dummyCode_RU,dummyCode_SI,APQ_ID

study variable=APQ_ID

t_threshold=2.5

tip=4

nonparametric=1

permutation=4000

length_threshold=20

select=

normalize_qa=1

no_tractogram=1

exclude_cb=1

fdr_threshold=2.5

running connectometry

read region from FreeSurferSeg:Left-Cerebellum-White-Matter

searching FreeSurferSeg from the atlas pool...

loading atlas FreeSurferSeg from the atlas list.

loading Left-Cerebellum-White-Matter from FreeSurferSeg atlas

read region from FreeSurferSeg:Left-Cerebellum-Cortex

searching FreeSurferSeg from the atlas pool...

loading atlas FreeSurferSeg from the atlas list.

loading Left-Cerebellum-Cortex from FreeSurferSeg atlas

read region from FreeSurferSeg:Right-Cerebellum-White-Matter

searching FreeSurferSeg from the atlas pool...

loading atlas FreeSurferSeg from the atlas list.

loading Right-Cerebellum-White-Matter from FreeSurferSeg atlas

read region from FreeSurferSeg:Right-Cerebellum-Cortex

searching FreeSurferSeg from the atlas pool...

loading atlas FreeSurferSeg from the atlas list.

loading Right-Cerebellum-Cortex from FreeSurferSeg atlas

total track=6 analysis restarted with higher seed count...

now running seed count=20000

total track=15 analysis restarted with higher seed count...

now running seed count=40000

total track=36 analysis restarted with higher seed count...

now running seed count=80000

total track=61 analysis restarted with higher seed count...

now running seed count=160000

.t2.nqa.fdr0.500000..cerebellum.pos_corr.tt.gz

output results

. A nonparametric Spearman partial correlation was used to derive the correlation, and the effect of Age, Age_Sq, Sex, dummyCode_CBIC, dummyCode_RU, and dummyCode_SI was removed using a multiple regression model. A total of 645 subjects were included in the analysis. A T-score threshold of 2.5 was assigned and tracked using a deterministic fiber tracking algorithm (Yeh et al. PLoS ONE 8(11): e80713, 2013) to obtain correlational tractography. The QA values were normalized. The tracks were filtered by topology-informed pruning (Yeh et al. Neurotherapeutics, 16(1), 52-58, 2019) with 4 iteration(s). An FDR threshold of 0.5 was used to select tracks. To estimate the false discovery rate, a total of 4000 randomized permutations were applied to the group label to obtain the null distribution of the track length. The tracking algorithm used cerebellum as the terminating region.

loading /opt/dsi-studio/dsi_studio_64/track/HCP1065.tt.gz

warping atlas tracks to subject space

trk files saved

Warning: --normalized_qa is not used. Please check command line syntax.

*** 

Thoughts on what I might be doing incorrectly?

To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/CAG_Qrt%2BFV10NHdrN6z6cpT9Hi%2B6dR%3DJ-YPb8PW4Noao4_O0ZVg%40mail.gmail.com.

[Frank Yeh]

In the new version I replace --normalized_qa by --normalize_qa.

Sorry for the change.

To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/CAN8Mcn8OizP6Ggy8NgNgg-nZJcOqBj-h7rA_59N6qA226S_D8w%40mail.gmail.com.

[Jamie Hanson]

Ah good to know, many thanks again for all the assistance!

(And apologies I didn't catch that right away!)

To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/CAG_QrtK8oYTZ_PZ3soGt%2B%2BMC5UWOyXDedft1iM_zrjQUzUzkJA%40mail.gmail.com.

[Jamie Hanson]

Drat, I'm still running into issues... additional thought on if I'm doing something wrong? 

jamielh@pfc:~/Volumes/Hanson/Public_Access_MRI/NKI_HealthyBrainNetwork/Neuroimaging/BIDS/Derived/dsi-studio/APQ$ singularity exec dsistudio_latest.sif dsi_studio --action=cnt --source=connectometry.qa.db.fib.gz --demo=APQ_combined.csv --variable_list=0,1,2,3,4,5,6 --voi=6 --permutation=1000 --fdr_threshold=2.5 --length_threshold=20 --t_threshold=2.5 --normalize_qa=1 --tip=4 --exclude_cb=1 --no_tractogram=1

DSI Studio Feb  5 2021, Fang-Cheng Yeh

QStandardPaths: XDG_RUNTIME_DIR points to non-existing path '/run/user/1002', please create it with 0700 permissions.

Starting GUI-based command line interface.

source=connectometry.qa.db.fib.gz

action=cnt

reading connectometry db

process 186637: D-Bus library appears to be incorrectly set up; failed to read machine uuid: UUID file '/etc/machine-id' should contain a hex string of length 32, not length 0, with no other text

See the manual page for dbus-uuidgen to correct this issue.

libGL error: No matching fbConfigs or visuals found

libGL error: failed to load driver: swrast

demo=APQ_combined.csv

variable_list=0,1,2,3,4,5,6

voi=6

variables=Age,Age_Sq,Sex,dummyCode_CBIC,dummyCode_RU,dummyCode_SI,APQ_ID

study variable=APQ_ID

t_threshold=2.5

tip=4

nonparametric=1

permutation=1000

total track=63 analysis restarted with higher seed count...

now running seed count=160000

.t2.nqa.fdr0.500000..cerebellum.pos_corr.tt.gz

output results

. A nonparametric Spearman partial correlation was used to derive the correlation, and the effect of Age, Age_Sq, Sex, dummyCode_CBIC, dummyCode_RU, and dummyCode_SI was removed using a multiple regression model. A total of 645 subjects were included in the analysis. A T-score threshold of 2.5 was assigned and tracked using a deterministic fiber tracking algorithm (Yeh et al. PLoS ONE 8(11): e80713, 2013) to obtain correlational tractography. The QA values were normalized. The tracks were filtered by topology-informed pruning (Yeh et al. Neurotherapeutics, 16(1), 52-58, 2019) with 4 iteration(s). An FDR threshold of 0.5 was used to select tracks. To estimate the false discovery rate, a total of 1000 randomized permutations were applied to the group label to obtain the null distribution of the track length. The tracking algorithm used cerebellum as the terminating region.

loading /opt/dsi-studio/dsi_studio_64/track/HCP1065.tt.gz

warping atlas tracks to subject space

trk files saved

jamielh@pfc:~/Volumes/Hanson/Public_Access_MRI/NKI_HealthyBrainNetwork/Neuroimaging/BIDS/Derived/dsi-studio/APQ$

jamielh@pfc:~/Volumes/Hanson/Public_Access_MRI/NKI_HealthyBrainNetwork/Neuroimaging/BIDS/Derived/dsi-studio/APQ$ ls

APQ_combined.csv  connectometry.qa.db.fib.gz  dsistudio_latest.sif  subject_list.txt

jamielh@pfc:~/Volumes/Hanson/Public_Access_MRI/NKI_HealthyBrainNetwork/Neuroimaging/BIDS/Derived/dsi-studio/APQ$

(basically, there's no files output)

[Frank Yeh]

You may send me the *.fib.db.gz file and the *.csv file, and I can
help test on my side and fix it.
Sorry for all the trouble.
Frank

> To view this discussion on the web visit https://groups.google.com/d/msgid/dsi-studio/CAN8Mcn-P57Du%2B50VRX6N1%2B6MY7M7Q-jGSpn7%3DbqQUcFA7BPgLw%40mail.gmail.com.