Is there a configure script for Cell Profiler?


We are trying to install Cell Profiler on our cluster, I was wondering if there was a configure script for Cell Profiler? There are already some of the dependencies installed on the cluster, and I want to use them but there is no way that I could specify it. Usually for linux programs, they should have a configure script that has many options that you could pass to it, but this one only has a Makefile which is usually the product of the configure script. So that makes it hard…

Any help of advice is greatly appreciated!!


Hi Matt,

We have wiki page on the subject of installing CP 2.0 on Linux, located here.

Hope that helps!

Hi Mark!!

Thank you for your reply. We are still having a bit of a problem. We have read the webpage that you referenced, but it didn’t really help. Your script builds everything from scratch. usually, there should be a configure script that could detect all dependencies, and you could provide the non-system-wide one with your own commandline parameters. for example, sth like

./configure --install-dir=/my/dir --with-python=/my/python/2.6.6 --with-atlas=/my/atlas/lib

things like that. right now, it just builds everything without trying to find system ones. This is really jamming things up, because we do have many different versions of the dependencies for Cell Profiler already on the cluster. If there is anything out there like that I am sure that it wouldn’t only benefit us.

Hi Matt,
Unfortunately, our build script is the result of quite a bit of trial and error, probably several people-months. CellProfiler, especially with its new Ilastik pixel classification, has a lot of dependencies. I am guessing that you’re right in wanting to use your system ATLAS / BLAS / LAPACK. However, some of the other dependencies are hard to navigate and we’ve found that we’ve had to build our own versions to get around incompatibilities - your system Pango is probably not going to work with the QT version we build, for instance.

You can use environment variables to tailor the build script. For instance, $ATLASVERSION and $ATLASPREFIX should let you choose the version and installation directory of your copy of Atlas.

I wish we had the resources to develop the typical configure / make / make install installation procedure. For now, this is the best effort of the Broad imaging group, our IT department and some of our valued users. It does take a long time to run. The output is stored in its own directory hierarchy and should not interfere with your system libraries in any way.