Most Massbank spectra dont resolve

[YasinEl]

Hello, it seems like the vast majority of Massbank accessions do not resolve.

Here is one example that does not work:

https://metabolomics-usi.gnps2.org/dashinterface/?usi=mzspec%3AMASSBANK%3A%3Aaccession%3AMSBNK-BAFG-CSL23111015084

here is one that works:

http://metabolomics-usi.gnps2.org/dashinterface?usi1=mzspec%3AMASSBANK%3A%3Aaccession%3AAC000204&width=10.0&height=6.0&mz_min=None&mz_max=None&max_intensity=125&annotate_precision=4&annotation_rotation=90&cosine=standard&fragment_mz_tolerance=0.1&grid=True&annotate_peaks=%5B%5B239.01060485839844%2C%20257.0210876464844%2C%20358.0841064453125%2C%20386.0790100097656%2C%20404.0895080566406%5D%2C%20%5B%5D%5D

Could be related to #204

Thanks!

[Michael-C-Strobel]

Hi Yasin, thanks for bringing this up.

I believe it's related to the non-standard dash used in the regex '-' vs '–' (0x002D vs 0x2013 in unicode). Will make a PR shortly.

[Michael-C-Strobel]

#206 should do it.

[YasinEl]

Thank you. So seems that solves another issue.

Specifically now mzspec:MASSBANK::accession:AC000204 (which worked before I created the issue) as well as mzspec:MASSBANK::accession:MSBNK-AAFC-AC000204 (which did not work before) resolve.

However, most massbank accessions still dont work in either format. For example: mzspec:MASSBANK::accession:MSBNK-Antwerp_Univ-AN118518

[Michael-C-Strobel]

I think we may need to file an issue over on MassBank to get some more details about their API. In both spectra, the Regex now correctly matches the ID. However, the API request for AN118518 does not return anything.

These are the requests made by the code.

Working:
https://massbank.us/rest/spectra/SM858102

Not Working:
https://massbank.us/rest/spectra/AN118518

[YasinEl]

thank you, will do that

[YasinEl]

linking it here MassBank/MassBank-web#409

[meier-rene]

Hi, I'm a maintainer of massbank.eu. We are closely related with the massbank.jp people and collect and curate the same collection of reference spectra called MassBank. We do this as the MassBank consortium(https://github.com/MassBank). MassBank exists since 2006.

We are not related to massbank.us which we usually call MoNA. massbank.us mirrors parts of our content. We did not upload that content at MoNA and we do not update it. This collection must be pretty old, because they do not contain the modification in the naming scheme, which we have done in mid 2022. It also only covers 72000 of our 124000 reference spectra.

We also realized that the original Massbank content is not properly covered by the spectrum resolver any more, but didn't find the time to investigate that issue further. Maybe we should take the opportunity to fix that.

First we need to find out which data is covered by the MASSBANK collection in the spectrum resolver. I always thought it is actually our data from MassBank, but given the content of this issue and after a short look in the code this seems to be not the case. I think its not correct. Was it always like this?

How can we continue from here? How do you want to handle MassBank and MoNA? What would you need to display MassBank content correctly?

[Michael-C-Strobel]

Hi Meier,
Thank you very much for the clarification. I agree, I think this would be a great opportunity to tackle this issue and get the new MassBank spectra resolving.

I think on the resolver-side, given that the ids will not overlap, we can hit a MassBank API for MassBank ids and the MoNA api for MoNA ids. In this way we don't introduce any new source accessions.

The most recent MassBank documentation I could find is here which is missing a way to request by MassBank ID. Is there an existing endpoint we can use?

It looks like this is the data for a spectrum: precursor_mz, charge, mz, intensity.

[meier-rene]

We have an REST API endpoint running under https://msbi.ipb-halle.de/MassBank3-api/v1. To fetch data for one record you can use eg https://msbi.ipb-halle.de/MassBank3-api/v1/records/MSBNK-IPB_Halle-PN000013, in general https://msbi.ipb-halle.de/MassBank3-api/v1/records/<accession id>.

The peaklist is found in ['peak'] ['peak'] ['values'].
precursor_mz would be ['mass_spectrometry']['focused_ion'] in the 'value' with 'subtag' 'PRECURSOR_M/Z'
For charge I dont know what is meant here. Charge of the molecular formula or precursor ion charge?

What other information would you need?

[Michael-C-Strobel]

Thanks for the API details. Regarding charge we're looking for the precursor ion charge, which I don't see a specific field for.

Taking a closer look at the parsing code, and a couple of the example identifiers here, it looks like most (in fact all that I checked) return a null value of zero.

I think we can leave it out for now and if it becomes available we can integrate it.

I believe this is all of the information we need to get this implemented. I'll come back and comment here if anything else comes up. Thanks again!

[YasinEl]

#207 should solve it. Big thank you to @Michael-C-Strobel