search

JSuryatenggara / ChIP-AP

An Integrated Analysis Pipeline for Unbiased ChIP seq Analysis
21 stars 11 forks source link

Several problems with ChIP-AP #5

Closed shelkmike closed 2 years ago

shelkmike commented 2 years ago

I encountered 5 problems with ChIP-AP 5.3:

1) In chipap_installer.py there is a line conda_dir = '{}/anaconda3'.format(conda_file.split('/anaconda3/')[0])
It will work only if the anaconda installation folder is named "anaconda3". Mine was /home/shelkmike/Work/Tools/Anaconda/ , so I had to edit chipap_installer.py

2) After running chipap_installer.py , the following error of anaconda occured: ResolvePackageNotFound: - r-base==3.2.2=0
I had to edit chipap_env_linux.yml and replace - r-base=3.2.2=0 with - r

3) On https://github.com/JSuryatenggara/ChIP-AP/wiki/Virtual-Machine-Installation-Guide there is a link to a virtual machine https://www.dropbox.com/s/4d3adb6ckof5rti/ChIP-AP_Ubuntu_LTS2004.vhd , but this file has been removed from Dropbox.

4) In chipap.py there is a line script_version = '5.1' even though the actual version is 5.3 . This makes chipap.py to print "Newer version of chipap.py is available on our github" when run.

5) When run, ChIP-AP printed The number of ChIP replicates (2) do not equal to control replicates (1). Do you accept this? (Y/N)
Please, make ChIP-AP capable of non-interactive behavior. When running on a computing cluster with a job scheduler like PBS, it is impossible to run interactive programs.

JSuryatenggara commented 2 years ago

Hello, thanks for your bug report and suggestions!

  1. Now the installer works when your anaconda installation parent folder is not named "anaconda3".
  2. Looks like new R packages are getting uploaded to anaconda such that the older ones are deleted. We have updated the .yaml file to fix the issue and to prevent such issue from happening again.
  3. New link to the dropbox has been added to the guides and wikis. Please note that these are older version of ChIP-AP. Virtual machine for the latest version of ChIP-AP (v5.4) is currently in progress and the link to it will be updated in around 2 days.
  4. Fixed.
  5. This user prompt is to make sure the user acknowledges that when the number of ChIP and control samples are unequal, ChIP-AP is going to calculate the peaks fold enrichment based on the merged ChIP and control samples (fcmerge mode on); and this prompt can be avoided by using the --fcmerge flag everytime the dataset has unequal number of ChIP and control samples. Nevertheless, after some consideration, we agreed to remove the prompt. It will now only print the message and proceed.

To get all the fixes in one go, please download and install the latest release (version 5.4)

Thanks again for your report and suggestion. Kindly let us know if you run across another issue.

shelkmike commented 2 years ago

Thank you.