CPC2 alternative?

[cschu]

The CPC2 download from http://cpc2.cbi.pku.edu.cn/data/CPC2-beta.tar.gz is currently unreachable, same for the main website. This tool causes issues with distributability since it is not installable from public sources (and needs to be patched in order to run with Python3). I do all of that in the singularity container, however, even the container cannot be built when the download source is not available.

I suggest:

• finding an alternative
• contacting the CPC2 lab and asking for permission to redistribute and patch the software with minos (there is no license in the code)
• ...

[gemygk]

References 13-20 from CPC2 paper are the ones we should try in that case

https://academic.oup.com/nar/article/45/W1/W12/3831091

[gemygk]

PLEK - https://pubmed.ncbi.nlm.nih.gov/25239089/

Cons: Python 2😢

https://sourceforge.net/projects/plek/files

[gemygk]

CPAT http://rna-cpat.sourceforge.net/#download-cpat

Pros: Python >3.5

[swarbred]

@cschu

I suggest we do both i.e.

We can look at other possible alternatives (continue to list here/summarise), and see if any seem to meet what we would require based on the tools description/paper/availability/requirements. Then we can see if we want to explore this further.

Go ahead and email cpc@mail.cbi.pku.edu.cn (you could also cc the corresponding authors from https://academic.oup.com/nar/article/45/W1/W12/3831091) and enquire about the current lack of availability and also describe how we want to make use of there tool to see what other solutions may be possible.

CC me in and I can chime in if required (and if we get a reply).

[swarbred]

Just to add, we also use this in a very limited way as currently we don't define any genes as lncRNA, it's there so we can exclude models with no clear coding potential and homology support from the protein_coding set. It's also used in the scoring (as one metric among many).

Any models marked as ncRNA in the mikado gff we automatically categorise as predicted gene low confidence #13 (comment)

[swarbred]

cnit
https://academic.oup.com/bib/article/20/5/1853/5047384
http://cnit.noncode.org/CNIT/download

[swarbred]

option ? https://www.biorxiv.org/content/10.1101/620880v1

[gemygk]

I will try installing the last two (CNIT and RNAsamba) and see how they perform.

[swarbred]

RNA samba was mentioned on a call today ensembl are looking at this (and CPC2)