Is there any way I could still use Metaboanalyst 5.0?

I want to use this version for the sake of uniformity. Is there any way I could access this version?


I agree with this sentiment. It’s unfortunate that when new versions of MetaboAnalyst are released, they replace prior versions with no overlap period to ensure functional continuity. Simply view the many postings in this forum lamenting the many faults with version 6.0, which have prevented users from generating new results or reproducing old results from earlier versions, thereby causing their research to be stymied or outright halted. Likewise, consider how new, but non-functional versions of MetaboAnalyst—which lack legacy support for prior versions—may contribute to the replication crisis plaguing all of science.

In fairness, version 4.0 can be downloaded and locally run using R, thereby allowing a user to “lock-in” a version and avoid forced updates. But, the local version does not contain the easy-to-use, user-friendly interface of the Web-based browser version, which frankly, is the entire appeal of using MetaboAnalyst.

Complaining about free software puts users in a difficult stance as nobody wants to appear ungrateful. I’m sure the designers of MetaboAnalyst and the moderators in this forum feel a certain amount of concern as they try to balance improvement with functionality. Perhaps having an overlap period or including legacy support may address those concerns.

Maintaining transparency and backward compatibility is among our top priority. MetaboAnalyst 6.0 has added new modules and enhanced current modules. The main change that I am aware of is the name mapping (from HMDB 4.0 to 5.0) which will affect pathway and enrichment analysis. There were some issues related to this change and most of them are due to cross mapping between KEGG and HMDB (as they are not synchronized 100%). We have manually fixed them - thanks to those who posted the details for us to reproduce.

I am curious about the differences or other inconvenience you have experienced. Note I have responded to several posts on this topic asking for more clarifications, none responded … Note we also published many tutorials and protocols based on the previous versions. If you have concerns with your data, you are welcome to use our own data and tutorials to show these cases.

Hello, I have a similar problem, with version 5, it identifies most of my metabolites, including phospholipids, for example
In my analysis I had to added few new characteristics and to generate a new pathway graph, it comes out completely different, it no longer recognizes most of the lipids and before the phospholipid route was the one with the greatest impact, now it does not even appear, and most of my data is still Phospholipids, only now they are all eliminated because the platform no longer recognizes the HMDB code.

Thanks for your reply! As I and others have mentioned, changing the shapes and colors in a PCA scores plot is not functioning and may cause the site to crash, which started to occur with version 6.0.

  1. (with detailed explanation and sample data, according the error-reporting guidelines of this forum)
    Colors & Shapes error with PCA
  2. PCA loading failed - can not modify the color of the sample
  3. Module for shape and color changing is not functional on V6.0

Please provide explicit example. We need to find out if this is due to database change or due to the matching algorithm.

OK, I can reproduce the issues related to shape and colors. It should be fixed (the server will be updated in 24 hours).

Two comments:

  1. The issues are more related to your parameter settings (i.e. duplicated colors and shapes) - it should not work in 5.0, as we don’t have that in mind until now. Such “creative use cases” are performed at your own risk since they are not documented on our tutorials.

  2. A lot of those issues are related to using MetaboAnalyst as GraphPad or graphical tools with extensive support for manual customization. These features are not research related and are challenging to maintain in long term. If these features are important for your work, please consider becoming subscribed / Pro users.

1 Like