Kegg Paywalled, What Do People Use?

Question

[Iddo]

Hi,

The updated KEGG FTP is now subscriber-only. My question is: what do people use instead? I am especially interested in prokaryotes (so, BioCyc maybe?) but is there a good updated replacement for KEGG?

Thanks,

Iddo

Answer 1

[Khader Shameer]

I am using the following resources for pathway analysis

• PANTHER (Protein ANalysis THrough Evolutionary Relationships) Classification System have a set of Pathways. Not many prokaryotic genomes are available at PANTHER. List of species here

• Reactome have orthology derived pathways for some prokaryotes. Extensive data in the level of molecular events available for human (Read about definition of Reactome molecular events here). See species comparison tool for pathways between human and any of the other species inferred from Reactome by orthology

• WikiPathways Not many prokaryotes there, but still useful for human/eukaryote-centric analysis

You may also start from PathGuide and see if there is any dedicated pathway resource for your taxa / species of interest.

[Joachim]

Thanks. I knew Reactome and WikiPathways, but PANTHER was new to me.

Answer 2

[Peter]

We're still using a copy of the final public release of KEGG before they closed off the FTP site - only a practical option in the short term sadly.

[Iddo]

Thanks. That's rather discouraging...

Answer 3

[Damian Kao]

I actually had to resort to web scraping their site. Wrote a script to get all the relevant html files with curl and then parsed the html file for the data. It's obviously not the best way to do it.

[mgalactus]

You could use the KEGG API, it should be more convenient and give you less headaches :)

Answer 4

[rama.gollapudi]

BioCyc/EcoCyc works well.

R

[Iddo]

Yep. But I'm looking for human too.

Answer 5

[vvanburen]

Hi Iddo, We are working on our own solution to this: http://vanburenlab.tamhsc.edu/cognoscente.html This is a biomolecular interaction knowledge base that provides a visualization of the query results.

The knowledge base is primarily built from EcoCyc, BIND, BioGrid, and HPRD interactions, which we we get via NCBI. We will keep up-to-date via NCBI, user submissions, and soon we will incorporate WikiPathways. One novelty of this tool is that queries are done across orthologs, and that information is preserved in the visualization. Another useful feature is that you can simultaneously query multiple genes/gene products.

Sadly, we don't have a data dump or any formal web services just yet, but that is our next order of business.

I look forward to any comments or suggestions: [e-mail]

Cheers, Vincent VanBuren Texas A&M HSC College of Medicine

Answer 6

[mgalactus]

It depends on what you need to do, but i've found that the KEGG API is a lot convenient and easy to use: the only (to my knowledge) thing you can't do is to map a set of proteins using blast, but you can use the KAAS web server (which can be used only through a browser, they told me they got no plans to implement an API for that)

Answer 7

[Charles Warden]

On one hand, I currently would get KEGG enrichment from Enrichr, and previous KEGG annotations should still be there (they list the year multiple times for many databases). It also looks like there are also still many free functions on the KEGG website, but maybe I need to look over everything more carefully.

On the other hand, I think having a variety of free programs available for various purposes is really important for research. While I would probably benefit from brainstorming alternative solutions for funding public goods (I would personally consider KEGG a public good, like most bioinformatics software / databases), I would strongly prefer seeing links for donations (kind of like some art museums) rather than paywalls or licenses. I would also prefer to pay for training, rather than paying for a subscription that limits access to the underlying information.

So, I am disappointed to hear that KEGG is adding some subscription options. For example, in the case of COHCAP, I would say I am providing a relatively direct solution to a problem (with straightforward statistical analysis that uses popular open-source languages / packages / libraries), and I think it would be better to have a freely available version with delayed/minimal support, rather than leaving the open-source license (although, of course, I would prefer an open-source solution that still provides timely support to users). Likewise, I think the inability to guarantee that KEGG (or COHCAP, etc.) can provide what is needed in any given situation means that there is some benefit to the lack of liability with an open-source license (in addition to helping justify an organization's non-profit status, if that is necessary), but I think considering something a public good should be the primary motivation. However, that is just my opinion as an individual :)