Getting started creating an EnzymeML file to store experimental data

[danolson1]

I am very interested in the EnzymeML project, but am having trouble getting started. My overall goal with EnzymeML is to store experimental data with relevant metadata and make it easy to share with collaborators who are doing mathematical modeling. I lead a group that primarily does experimental work, and have some experience with Python programming.

Here's an example of what I'm trying to do:
I collected data for the formate dehydrogenase enzyme in a 384 well plate with varying concentrations of formate and NAD+. I would like to create an EnzymeML file that stores the following:

• Metadata about the experiment (experimenter, date collected, instrument used)
• The overall reaction conditions (enzyme, reaction, buffer composition, pH, temperature)
• Timecourse data for each well

Here are some questions I've come across:

1. For a multi-well plate, is the vessel object the whole plate or an individual well of the plate? Or is this what the measurement group_id is for?
2. What is the prefered way to specify buffer composition?
3. How do I specify the ModifierRole for a species, for example to indicate that it is part of the buffer composition?
4. Is there a prefered way to indicate the reagent that was used to start the reaction?

Once I get this working, I'd be happy to share it as an example if other people are interested.

Regards,
Dan

[danolson1]

Also, what is the right place to start in the documentation? I did a lot of reading, but there were a lot of dead ends or out of date information:

• I spent a bit of time with the excel template that was mentioned in one of the publications, but that doesn't seem to be supported in v2 of the software
• I looked through the PyEnzyme documentation, but it is only on version 1.1.4
• I looked at the PyEnzyme examples, but the examples seemed more focused on modeling than packaging of real-world data.
• The EnzymeML training course looked promising, but only Session 1 and Session 2 have any information.

[JR-1991]

@danolson1 apologies for the late reply, I was on vacation when opened and it unfortunately has gone under my radar.

First of all, we are happy about your interest in the EnzymeML format! Happy to answer above questions:

---

For a multi-well plate, is the vessel object the whole plate or an individual well of the plate? Or is this what the measurement group_id is for?

Yes, the Vessel object is the right place to start. You can define each well in this object to keep track of all the wells, but since the main usage of this object is to turn possible mass measurements into concentrations, defining the plate once with the per-well volume is sufficient. My colleague @haeussma has developed a library MTPHandler to parse MTP-based measurements into EnzymeML.

---

What is the prefered way to specify buffer composition?

You can define the individual components as SmallMolecule and for convenience reference each of these in a Complex object, which can then be used in a Reaction object. Below is an example code that demonstrates how it can be done:

import pyenzyme as pe

enzmldoc = pe.EnzymeMLDocument(name="Buffer Demo")

# Mock small molecules for now
substrate = pe.SmallMolecule(name="Substrate", id="S")
product = pe.SmallMolecule(name="Product", id="P")

# Fetch the buffer components from ChEBI
tris = pe.fetch_chebi("CHEBI:9754")
edta = pe.fetch_chebi("CHEBI:64734")

# Add the small molecules to the document
enzmldoc.small_molecules.extend([tris, edta, substrate, product])

# Create a buffer complex
buffer = pe.Complex(id="C1", name="Buffer", participants=[tris.id, edta.id])
enzmldoc.complexes.append(buffer)

reaction = pe.Reaction(id="R1", name="Reaction")
reaction.add_to_reactants(substrate.id, stoichiometry=1)
reaction.add_to_products(product.id, stoichiometry=1)
reaction.add_to_modifiers(buffer.id, role=pe.ModifierRole.BUFFER)

# Save the document
pe.write_enzymeml(enzmldoc, "buffer_demo.json")

---

How do I specify the ModifierRole for a species, for example to indicate that it is part of the buffer composition?

As above, when using the Complex object, you can describe the composition and in a Reaction indicate its role via a modifier of the reaction. We have settled for this design to provide flexibility, as it can theoretically happen that roles of a molecule can differ in between reactions.

---

Is there a prefered way to indicate the reagent that was used to start the reaction?

There is currently no explicit way of describing which reagent has started the reaction, but we are happy to work on a way to extend the data model to accommodate for this common case.

---

Also, what is the right place to start in the documentation? I did a lot of reading, but there were a lot of dead ends or out of date information

We are currently in the process of migrating to the new version of EnzymeML and updating the documentation due to a few changes in the code base.

Regarding the EnzymeML spreadsheet, we have deprecated it for several reasons. Firstly, the parsing has been prone to errors due to the varying contents of cells and the structural limitations of Excel. Secondly, we had to incorporate Visual Basic for validation, which typically triggers virus detection when sharing the spreadsheet via email.

To address these issues, we have developed the EnzymeML Suite, a graphical user interface for EnzymeML. This suite offers a more streamlined experience with enhanced functionality compared to the spreadsheet. While its primary use is for non-technical users, it provides a direct entry point into Python, allowing users to reuse the data tracked in the Suite.

At the bottom of this page, you can download the current build for your operating system:

https://github.com/EnzymeML/enzymeml-suite/actions/runs/18114359449

---

The EnzymeML training course looked promising, but only Session 1 and Session 2 have any information.

We’ve planned an on-site training course for mid-January. If you’d like, I can send you the announcement via email (jan.range@simtech.uni-stuttgart.de). Additionally, if you prefer, we can schedule an online meeting, and I’ll guide you through our typical workflows with EnzymeML. Of course, I’ll also provide support for your own application.

[danolson1]

Thanks. Thats very helpful. Adding your buffer example (above) to the examples folder might be helpful for other people. In fact, adding several more examples to the examples folder would be helpful, and could be implemented even before the documentation update is done.

Yes, please add me to the email list for updates and training.

[danolson1]

I have played around with the EnzymeML suite. Initially, I thought it might be useful for technicians or students to use, but I haven't found a good use case yet. For members of my group with low computer literacy, training them to use the EnzymeML suite is too much work. It's easier for me to process the data myself in Python. For members of my group with medium to high computer literacy, it's easier to teach them to use the Python interface than to teach them 2 different tools. Improvements in AI tools for writing code has made the Python route even easier.

[JR-1991]

Okay, that makes sense. Fortunately, these tools are becoming increasingly valuable. I’ve added you to the EnzymeML mailing list and will follow up with the announcement of the next EnzymeML Training Course.

We’ll create a couple more examples that aren’t focused on modeling, but rather on data acquisition. I’ll open a pull request for these and will notify you here.

[danolson1]

I'm happy to send you the files I've been working on, both to get some feedback on how I've been doing this, and for you to use (or modify) as example files.

Some of them are on this github repository. I can check with Janis to see if she can invite you to join, or I can just email you the files.