The task is to select “good cells” based on cell size. However, due to variance in mean cell sizes between cell lines in different wells, we would like to filter out cells not on absolute numbers but percents from maximun or mean area of the cells in one well. The “select population” building block should be used.
Solution:
Getting the following error from Columbus 2.6 Insight:
---------------
Login Failure
---------------
Please check your user name and/or pasword or try again later.
---------------
OK
---------------
while the user name and passord are correct.
1) The Acapella disk cache resides in the following location: /var/lib/acapella/.imacro/diskcache Its fine to delete the contents if there are no running jobs on the server, switch to the Columbus job status page to check. In theory the max you would ever recover is 4GB as Acapella monitors the size of the folder to ensure it is kept below that level.
To remove:
$ sudo /etc/init.d/columbus stop
$ sudo rm -rf /var/lib/acapella/.imacro/diskcache/*
$ sudo /etc/init.d/columbus start
---------------------------------------
2) OMERO stores it's own database logs but they aren't used by Columbus. The logs which is likely to take up most of the space and which definitely aren't utilized by Columbus are found under the 'pg_log', this can be safely removed. It can be found in the following location:
/var/lib/pgsql/data/pg_log
Removing unused database logs:
$ sudo /etc/init.d/columbus stop
$ sudo rm -rf /var/lib/pgsql/data/pg_log/*
$ sudo /etc/init.d/columbus start
---------------------------------------
3) For any large database it's possible to reduce the database size by truncating the eventlog table, which is a fairly large table and contains information that Columbus does not use in any way.
##NOTE## Only perform these steps once you have confirmed that you have a recent database dump stored in /OMERO/OMERO4_4/db_backup. This backup can be used to recover the eventlog table to the database in the event of an error.
The command line workflow would be as follows
$ sudo su columbus
$ psql omero4_4
truncate eventlog;
\q
In case of an error, its possible to restore the original table from a complete dump like this:
$ pg_restore -v -c -t eventlog -d omero4_4 omero4_4-dateofmostrecentdump.pg_dbdump
There are quite a few dependencies, it's possible to find those using rpm, the following gives the packages we install with Columbus: $ rpm -qa | grep Columbus $ rpm -qa | grep Acapella Then on each of those outputs you'd run the following: $ rpm -q -R e.g. $ rpm -q -R Columbus-db I did this previously for Columbus 2.3/2.4 and the list at that time looked something like this. There have been several patch updates since then so it's a little out of date but the items listed should be the same for the latest release (just versions will differ slightly): glibc > 2.4 libjpeg libpng libtiff ncurses zlib freetype tomcat5 libICE.so.6()(64bit) libMagickCore.so.5()(64bit) libMagickWand.so.5()(64bit) libSM.so.6()(64bit) libX11.so.6()(64bit) libXext.so.6()(64bit) libXt.so.6()(64bit) libbz2.so.1()(64bit) libc.so.6(64bit) libdl.so.2()(64bit) libfontconfig.so.1()(64bit) libfreetype.so.6()(64bit) libgcc_s.so.1 libgdk_pixbuf-2.0.so.0()(64bit) libglib-2.0.so.0()(64bit) libgmodule-2.0.so.0()(64bit) libgobject-2.0.so.0()(64bit) libgomp.so.1()(64bit) libgs.so.8()(64bit) libjpeg.so.62()(64bit) liblcms.so.1()(64bit) libm.so.6()(64bit) libpng12.so.0()(64bit) libpthread.so.0()(64bit) librsvg-2.so.2()(64bit) libtiff.so.3()(64bit) libxml2.so.2()(64bit) ice-python > 3.3 ice-servers > 3.3 jre = 1.6.0_29 python > 2.4 libcrypt.so.1()(64bit) postgresql > 8.1 postgresql-server > 8.1 python-hashlib python-sqlite2
Hi,
The Columbus image analysis output files have special characters such as "Image Region Area [µm²] - Sum per Well." These special characters are problematic for LIMS and informatics software. Is there a way to exclude special characters from image analysis output files?
Thanks!
Christina
Hi Everyone,
We are very excited to announce the release of ColumbusTM 2.7!
Highlights of ColumbusTM 2.7 release include the following:
Your Columbus server administrator or manager should have received a notification with all the necessary information and files to upgrade your server. Please contact support (informatics.support@perkinelmer.com) if you have not received this information from PerkinElmer@flexnetoperations.com. Also, I have included the contents of our release notes below.
Seungtaek Lee | Informatics – Product Manager
PerkinElmer | For the Better
HUMAN HEALTH | ENVIRONMENTAL HEALTH
Supported Browsers
- Columbus fully supports:
- Firefox
- Google Chrome
- Internet Explorer 11
- Safari 5 or above
- Columbus is compatible with:
- Internet Explorer 10
- Other browsers and browser versions may function but compatibility is not guaranteed.
Known Issues
- Acapella scripts written in previous versions may not be compatible with Acapella 4.0 and Columbus 2.5. For information on how to convert old scripts to the new format refer to Technical Note 476.
- In Safari web browser, files downloaded from the links in the in the ‘Published Info’ tab may have extension .txt.csv instead of just .txt. This is automatically done by the Safari browser when it detects a file as ‘text/csv’.
Columbus 2.7.0
- Added automatic triggering of batch image analysis upon Harmony plate import by selecting a default analysis sequence for the screen (COL-1128)
- Added REST API to enable client applications to search and browse the Columbus database for direct data and metadata download (COL-975)
- Added ability to publish Screens and Plates in Columbus, in order to search, preview, and downloaded directly in High Content Profiler and TIBCO Spotfire (COL-985; COL-1085)
- Added ability to directly downloaded and merged or linked cell and well level results and assay definitions (plate layout) within High Content Profiler and TIBCO Spotfire (COL-1085)
- Added XYZ-View for z-stacks, it shows an xy-, xz- and yz-section through an image stack of a thick sample (e.g. microtissues) to allow for better interrogation of 3D image sets (COL-1156)
- Added support for Assay Specific Building Blocks in analysis sequences (COL-1157)
- Enhanced "Defined Results" Building Block for easier navigation and selection of features, including all features (COL-1159)
- The Secondary Analysis has been deprecated and might be removed in the future. It is recommended to use the High Content Profiler for the secondary analysis.
- Fixed pasting of data from Excel into concentration layer in the Assay Definition Editor (COL-1133)
- Fixed import configuration where folder names containing certain characters were not accepted (COL-274)
- Fixed Celery message queue hanging (COL-1420)
- Fix log file rotation for Columbus webapp logging (COL-1192)
- Improved default export folder name setting (COL-1444)
- Improved documentation (COL-1252, COL-1258, COL-1102, COL-405)
I was wondering if it's possible to import plates into Columbus from directories that are local to the Columbus server, which should then circumvent the Columbus Helper.
I am fairly certain that I was able to do this in the past by specifying a local directory with the syntax 'local:///path/...' instead of 'client://C:/...', but recent attempts in Columbus v2.6 have not been successful.
Is this possible? And if so, what is the proper syntax?
Thanks.
We are having trouble deleting plates and measurements from our Columbus server. Specifically, we have a plate that was read ~50 times for a stress test, so in our Columbus structure, we see a single plate with ~50 measurements. We are unable to delete either a single measurement or the parent plate folder. Other plates from the same user can be deleted without any issues. Are there any other options for deleting this data, perhaps at the command line?
Thanks in advance for your help.
Hello,
Is it possible to allow an administrator (root or otherwise) to save analyses and run batch analyses on other user's data once LDAP authentication is in place?
Thanks,
Chris
There is a known issue connecting to Columbus if the client PC running the clients is utilizing Java 8 - see forum post here:
http://forums.cambridgesoft.com/messageview.aspx?catid=55&threadid=3188&enterthread=y
However it has recently come to light that there is also a separate java compatibility issue which might also cause problems when attempting to connect to Columbus via the OMERO clients.
SSL ciphers are not supported by Java 7u85:
http://blog.openmicroscopy.org/tech-issues/2015/07/21/java-issue/
It is necessary to patch the /usr/local/PerkinElmerCTG/Columbus/etc/grid/templates.xml file as per the note above.
Glencoe have also provided a more up to date version of the OMERO clients available here:
Windows:
https://perkinelmer.box.com/s/7wya1smajvn0c3v61f9nurg6oqhvtdjj
Mac:
https://perkinelmer.box.com/s/ougp1zc9fy9y4xbcj5k5gyisgp7x4sgn
----------------------
- PerkinElmer Technical Support
Can we upgrade directly from v2.6 to v2.7.1, or do we need to upgrade through v2.7?
Thanks,
Chris
Highlights of Columbus 2.7.1 release include the following (full details are available at the bottom of this message as well as in the release notes):
Details of new features and bug fixes:
-Added ability to import Celigo datasets (COL-1466)
-Added ability to import Ensight/Kaleido 1.2 datasets (COL-1467)
-Added ability to import Vectra datasets (COL-1589)
-Added ability to import MetaSystems datasets (COL-1468)
-Added ability to import Zeiss datasets (COL-1469)
-Added ability to import Biotek datasets (COL-1496)
-Added ability to filter tables showing published status of screens and plates (COL-1276)
-Improved metadata capture for Arrayscan datasets (COL-1610)
-Improved metadata support for Columbus exported and transferred datasets, flat-field correction and orientation matrix information is maintained (COL-1424, COL-1503, COL-1504)
-Improved user interface, removed inconsistencies (COl-1227, COL-1312, COL-1407)
-Improved default memory settings, increased maximum Java VM heap space to avoid out of memory errors (COL-1507)
-Improved publishing of assay definition layouts, concentration values and unit strings are now in separate columns in the csv-text file download to be easily consumed by the High-Content-Profiler (COL-1637)
-Security enhancements for user authentication and file uploads (COL-1634, COL-1594)
-Fixed problem when importing timeseries Cell Voyager 7000 datasets where images where created out of sequence (COL-19)
-Fixed problem with email validations on user profile page (COL-952)
-Fixed problem when importing ScanR datasets with similar channel names which caused an import failure (COL-1285)
-Fixed version number of SOAP web-service to match Columbus version (COL-1382)
-Fixed problem with printing an image analysis sequence including a "Filter Image" building block (COL-1425)
-Fixed problem where images from Columbus when displayed Spotfire did not have the same channel coloring (COL-1456)
-Fixed problem with automatic log file cleanup, where old PostgreSQL log files have not been removed (COL- 1491)
-Fixed problem importing analysis scripts with too long file names (COL-1578)
-Fixed problem where well result tables in the secondary analysis page and error logs could not be downloaded (COL-1629)
If, after upgrading Columbus you have trouble connecting or transferring data from Harmony to Columbus and receive the following error:
ERROR: 8/8/2016 9:31 AM Transfer to Columbus Cannot create Columbus job.
Error code: 5003, error text: 'The request failed because the authentication token was invalid.'
it is likely that the Columbus account needs resetting in Harmony. Simply re-entering the Columbus user credentials via Harmony is usually sufficient but on occasions it is necessary to use the following workflow to create a new authentication token:
1) In Harmony select 'Settings' ...'User Accounts' ...'Manage Columbus Account'
2) Leave the Columbus URL the same but temporarily change the user account and password to something completely different, i.e. use the details of a different columbus user account e.g the 'columbus' or 'root' account...then select 'OK'.
3) Select 'Manage Columbus Account' again
4) now enter the correct user account and password for your columbus account ...select 'OK'
The act of re-entering the password in this way will reset the authentication token and should resolve the problem.
RH
I'm currently unable to save an analysis file from the Image Analysis tab to a database folder in my own account. I'm getting the following error:
Columbus could not handle the request:
OMERO/Ice exception: omero::SecurityViolation
/OMERO/PerkinElmerCTG/Acapella-4.0.2/AcapellaResources/AcapellaColumbusWebapp/ProcLib/columbus/util_treeview.proc(371) [Columbus::nav_AALSaveAnalysisAs]: /OMERO/PerkinElmerCTG/Acapella-4.0.2/AcapellaResources/AcapellaColumbusWebapp/ProcLib/columbus/nav_analysis.proc(365) [columbus::SetAnalysis]: OMERO/Ice exception: omero::SecurityViolation
I'm able to download the .aas file and run batch analysis; the analysis runs but after the last well, the results don't save to the database, which makes me think there might be a write restriction on a particular directory.
Please let me know if you have any suggestions/workarounds.
Thanks,
Chris
Targeting Log Files
Columbus:
The Columbus log files are found under /var/log/columbus
If the PostreSQL database fails to initialize, check the content of:
/var/log/columbus/db/master.err
For all other PostreSQL database issues, check the content of:
/var/log/columbus/db/Blitz-0.log
If the Acapella server fails to initialize, check the content of:
/var/log/columbus/acc/acapella.log
For all other Acapella issues, check the content of:
/var/log/columbus/acc/acc.log
If the Web Server is not running, check the content of:
/var/log/columbus/web/columbus.log
/var/log/columbus/nginx/error.log
Tomcat:
The Tomcat log files are found under /var/log/tomcat6
If the Tomcat startup fails or the Columbus web service is not accessible, check the content of:
/var/log/tomcat6/catalina.out
If the Web Service is accessible yet clients receive an error from the CCInterface2 web service, check the content of:
/var/log/tomcat6/CCInterface2.log
Extracting Log Files
Columbus:
A useful sub-section of log files can be easily extracted via a web browser using the following URL:
http:///help/error/all_logs.zip
Alternatively, the full log set can be compressed and bundled into the /tmp directory on the server by issuing:
$ zip –r /tmp/columbus_logs.zip /var/log/columbus
The compressed columbus_logs.zip file can be exported from the Columbus servers /tmp directory using WinSCP
Tomcat:
The full log set can be compressed and bundled into the /tmp directory by issuing:
$ zip –r /tmp/tomcat_logs.zip /var/log/tomcat6
The compressed tomcat_logs.zip file can then be exported from the Columbus servers /tmp directory using WinSCP
I am trying to install Columbus on CentOS 6. Everything seems fine however I am having issues with Postgresql. This is the error message that I get:
Installing : Columbus-db-2.5.0.120577-gen.x86_64 5/162
Non-fatal POSTIN scriptlet failure in rpm package Columbus-db-2.5.0.120577-gen.x86_64
postmaster (pid 775) is running...
creating postgres user columbus
Stopping postgresql service: [ OK ]
Starting postgresql service: [FAILED]
Installation aborted! Could not restart postgres database!
warning: %post(Columbus-db-2.5.0.120577-gen.x86_64) scriptlet failed, exit status 1
AND this:
The trick is to use the select region BB. You select the border and calculate the area of that border.
Then use calculate properties to get the correct units [µm].
Stack Processing (fixed, string) : Individual Planes (, string) Flatfield Correction (fixed, string) : None (, string) Stack Processing (fixed, string) : Individual Planes (, string) Flatfield Correction (fixed, string) : None (, string) Channel (fixed, string) : DPC (, menu) Select channel for cell detection ROI (fixed, string) : None (, menu) Select population for ROI Method (fixed, string) : B (, menu) Select method for cell detection Common Threshold (fixed, string) : 0.4 (default, numeric) Controls common threshold calculation. Adjust by the tuning illustration Initial Mask. Area (fixed, string) : > (fixed, string) 100 (default, numeric) µm²(fixed, string) Minimum allowed area of cell. All objects with area less than the limit are discarded as artifacts. Split Factor (fixed, string) : 7.0 (default, numeric) The Split Factor influence on splitting of clump cells to the individual objects. Individual Threshold (fixed, string) : 0.4 (default, numeric) The parameter adjusts the object borders, i.e. the detected cells could appear larger or smaller. Contrast (fixed, string) : > (fixed, string) 0.1 (default, numeric) Defines the minimum allowed contrast of cell. All objects with contrast less than the limit are discarded as artifacts. Adjust by the tuning illustration Low Contrast Cells. Output Population (fixed, string) : Cells (, string) Name of the output population Channel (fixed, string) : DPC (, menu) Select channel for cell detection ROI (fixed, string) : None (, menu) Select population for ROI Output Population (fixed, string) : Cells (, string) Name of the output population Population (fixed, string) : Cells (, menu) Select population Region (fixed, string) : Cell (, menu) Select region by which a new region should be created. Method (fixed, string) : Standard (, menu) Select method by which a new region should be created Border (fixed, string) Turn the border calculation ON/OFF Output Regions (fixed, string) : Cell (, string) Common prefix for region names Population (fixed, string) : Cells (, menu) Select population Region (fixed, string) : Cell (, menu) Select region by which a new region should be created. Output Regions (fixed, string) : Cell (, string) Common prefix for region names Population (fixed, string) : Cells (, menu) Select population Region (fixed, string) : Cell Border (, menu) Select region which should be characterized. Method (fixed, string) : Standard (, menu) Select a method Area (fixed, string) Turn the area calculation ON/OFFUnit for area property Output Properties (fixed, string) : Cell Border (, string) Common prefix for the property names Population (fixed, string) : Cells (, menu) Select population Region (fixed, string) : Cell Border (, menu) Select region which should be characterized. Output Properties (fixed, string) : Cell Border (, string) Common prefix for the property names Population (fixed, string) : Cells (, menu) Select population, for which a custom defined property should be calculated. Method (fixed, string) : By Formula (, menu) Select a method Formula (fixed, string) : A (, string) Calculate a new property with formula in style 100*(A-B)/C and thereupon specify below the corresponding variables A, B etc. Please use only single characters (A, B, ..,Y). Variable A (fixed, string) : Cell Border Area [µm²] (, menu) Select a property to which the variable A should correspond to Output Property (fixed, string) : Perimeter [µm] (, string) Name of the calculated property Population (fixed, string) : Cells (, menu) Select population, for which a custom defined property should be calculated. Output Property (fixed, string) : Perimeter [µm] (, string) Name of the calculated property
Analysis Sequence "[New Analysis]"
Analysis Sequence "[New Analysis]"
Input Image
Input Image
Input
Find Cells
Method
Find Cells
Input
Output
Select Region
Method
Select Region
Input
Output
Calculate Morphology Properties
Method
Calculate Morphology Properties
Input
Output
Calculate Properties
Method
Calculate Properties
Input
Output
Define Results
Define Results
Input
Acapella version: 4.1.2.118496. Timestamp: 2016-10-06 10:12:51 +0200.
Hi, I am trying to use Columbus' "Find Texture" module by training, but I must be missing something. I setup the building block, and then click on the "Train..." button , but nothing happens when I try and then click on the image. Is there some other trick besides left-click? This just works as a pixel classifier, right? I don't necessarily need to use upstream objects or measurements, do I?
Screenshot of workflow: https://cl.ly/3O1T170l0J1I
Thanks for any advice.
David
If I output texture features of the cytoplasm, there is no common feature to equate it to the spots found in the same cell. Spots are output as their own population with no common, cell ID, X/Y coordinate etc to the nuclei population even though they were found using the Nuclei population. One can not then use the spot and cytoplasm texture or STAR features together for analysis.
----------------------
This may occur following Columbus updates due to internal software restructuring. To run the analysis sequence in a batch analysis, follow the instructions given in the error message displayed in the job status page:
1) Load the sequence on the Image Analysis screen
2) Click on the "Define Results" building block to run through the sequence once
3) Save the sequence
4) Repeat the Batch Analysis using the new sequence
Refreshing the sequence in this way will ensure that the analysis sequence is compatible with the current version of Columbus.