How to install GenomeWorks on local machine?

[kountaydwivedi]

Hello everyone,
I am a research scholar, and I need to test GenomeWorks on sample data. Could anyone please explain how to install and run on my system?
System details:
UBUNTU 18.0
6GB CUDA -enabled NVIDIA GTX 1660Ti
Intel core-i7 9th Gen
16GB RAM

Thanks in advance.

[tijyojwad]

Hi @kountaydwivedi - the setup instructions for GenomeWorks are described in https://github.com/clara-parabricks/GenomeWorks#genomeworks-setup .

Are you interested in using the C++ APIs or python APIs?

[kountaydwivedi]

Thanks @tijyojwad for your response.
I have installed GenomeWorks as per the setup instructions. But I'm unable to execute any command or tool in the terminal, unlike the GATK4 application, in which I could easily access each commands after installation. For eg:
I am trying to access-
$~ cudamapper
it is showing "command not found"

Also, I am going to use python APIs.

[kountaydwivedi]

Is is possible to use GenomeWorks with Anaconda on Windows/Ubuntu?

[tijyojwad]

Hi @kountaydwivedi - we don't install our binaries/libraries into any default system location. The easiest to add the binary to your path would be (on Ubuntu e.g.) export PATH=$PATH:<build/install/bin>.

We support Ubuntu but not windows. And an Anaconda package is on our roadmap but we don't have one generated yet. We plan to add that soon!

For the python API you would need to follow the instructions under pygenomeworks folder to install from source, or you can simply install from PyPi - e.g. https://pypi.org/project/genomeworks-cuda-10-1/ for CUDA-10.1

[kountaydwivedi]

Thanks @tijyojwad for your response. I would try to solve the issue. If there would be a problem, I'll reply to you.
Thank you once again.

[ahehn-nv]

Hi @kountaydwivedi could you solve the problem? If so, we would close the issue. If not, happy to help again.

[PMuchina]

Am having trouble installing Genome works on a Jetson nano... Anyone with tips on how to go about it?

[PMuchina]

Am trying to setup genome works on the jetson nano ... but it keeps exiting at the 'make -j install' stage with an error

[ahehn-nv]

Hi @PMuchina.

First of all, running GenomeWorks on NVIDIA Jetson Nano is not supported, since GenomeWorks requires a GPU with CUDA compute capability >= 6.0 and NVIDIA Jetson Nano has compute capability 5.3.

However, you might be able to try GenomeWorks on NVIDIA Jetson Nano nevertheless, if you modify the CMakeLists.txt files and replace all occurrences of "-arch=compute_60" with "-arch=compute_53". (If I remember correctly, the only occurrences are in cudaaligner/CMakeLists.txt and cudapoa/CMakeLists.txt.) Please note that this has not been tested at all, is not supported officially, and may show poor performance.

Unfortunately, it is hard to understand what went wrong in your case specifically, since you did not mention what error you got during the build. I might be able to help if you post the message.

[PMuchina]

Thank you..... Finally I was able to install it..... My interest was cudamapper and I was able to run it successfully and get the expected results.....

…

On Monday, 8 February 2021, ahehn-nv ***@***.***> wrote: Hi @PMuchina <https://github.com/PMuchina>. First of all, running GenomeWorks on NVIDIA Jetson Nano is not supported, since GenomeWorks requires a GPU with CUDA compute capability >= 6.0 and NVIDIA Jetson Nano has compute capability 5.3. However, you might be able to try GenomeWorks on NVIDIA Jetson Nano nevertheless, if you modify the CMakeLists.txt files and replace all occurrences of "-arch=compute_60" with "-arch=compute_53". (If I remember correctly, the only occurrences are in cudaaligner/CMakeLists.txt and cudapoa/CMakeLists.txt.) Please note that this has *not* been tested at all, is *not* supported officially, and *may* show poor performance. Unfortunately, it is hard to understand what went wrong in your case specifically, since you did not mention what error you got during the build. I might be able to help if you post the message. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#593 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ANGRD6D4KGHMMLI5E7UPEF3S6BAPTANCNFSM4TCQWEGA> .