1. Philosophy & Motivation

At OpenADMET, “Open” is our first name. It is not just a moniker; it is our fundamental operating philosophy. We believe that the future of ADMET modeling relies on breaking down data silos, encouraging radical transparency, and fostering an inclusive, collaborative global community. We are committed to policies that foster rapid innovation and leverage the pace of scientific discovery we are currently experiencing.

• Data is Better Shared: Data is most valuable in the hands of the community, not locked away. Our blind challenges are one of the main ways we share it. We aim to run them as often as possible, on new and stimulating datasets, with all data released openly once each challenge closes.
• Respectful Debate and Discourse: Good science needs disagreement. We welcome challenges to our data, methods, and conclusions, including from people who disagree with us. We ask that all debate focus on the evidence, not the person, and that everyone engaging with OpenADMET follow our Code of Conduct.
• Continuous Improvement: We are not perfect, and we won’t always get things right. We treat these principles as a living commitment. When we fall short, we will say so openly and explain what we are changing. We welcome feedback from the community on how we can do better, through Discord, GitHub, or directly to the team, and we will review and update these policies regularly.

2. Operating Practices & Community Engagement

OpenADMET succeeds when the community succeeds. We are dedicated to building a vibrant, supportive, and collaborative ecosystem where researchers from academia, industry, and government can intersect. We actively express this commitment through multiple touchpoints:

• Interactive Webinars: We host regular, accessible educational webinars to share new datasets, walk through open-source code implementations, and highlight breakthrough methodologies.
• Active Discord Community: We maintain an open, real-time community chat platform on Discord. This serves as our digital watercooler: a place to troubleshoot code, discuss data anomalies, coordinate challenges, and foster informal peer-to-peer mentorship.
• Conferences and Convenings: We bring the community together in person through targeted conferences and consortium meetings. These events are designed to share cutting-edge science, map out future blind challenges, and strengthen the human connections that power open-source collaboration.

3. Publishing Practices & Open Outputs

Traditional methods of sharing scientific advances lag far behind the current pace of innovation in ML/AI. To keep up with and benefit from this speed, OpenADMET is committed to making all of our scientific outputs readily and rapidly available.

Open Source Software & Open Publishing

• Code First: All software tools, benchmarking pipelines, machine learning models, and infrastructure developed by OpenADMET will be released under permissive open-source licenses (MIT or Apache 2.0) on open repositories such as GitHub.
• Open Access: Scholarly outputs will be published rapidly as preprints (e.g., on bioRxiv or arXiv), appear in open access and not-for-profit venues, and be licensed with the Creative Commons Attribution (CC BY 4.0). OpenADMET will also respect all funder publication policies with regard to open-access archives and publication venues.
• Rapid Scientific Communication via Blog Posts: Technical blog posts serve as our primary means of rapid scientific communication. When we discover an interesting trend, uncover a data nuance, or develop a new workflow, we publish it to the community immediately under a CC BY 4.0 license. Blog posts are hosted on ghost.io and archived on Zenodo.
• Executable Science: To ensure these insights are immediately useful and reproducible, our scientific blog posts are accompanied by open-source software implementations, Jupyter/Marimo notebooks, or code/dataset archives hosted on GitHub, Zenodo, or Hugging Face.

Open Data Commitment

Data quality and accessibility are the bedrock of reliable machine learning models. We aim to create a frictionless data ecosystem where interested parties can use our data with ease under CC BY 4.0 licensing.

• Chemical Structures: All chemical structures used in our datasets, training sets, and evaluation suites will be fully disclosed using clear, standard machine-readable formats (such as explicit, standardized SMILES strings). There will be no obscured or proprietary structures.
• Assay Data: All experimental endpoints, kinetic data, and biological assay results will be made fully public under CC BY 4.0, accompanied by metadata describing experimental conditions. Where applicable, this data will also be deposited in open repositories such as ChEMBL, Zenodo, and the PDB.
• Instrument-Level Raw Data: True reproducibility requires knowing exactly how the data was generated. Wherever feasible, we commit to releasing raw data directly from the instrument level (e.g., raw LC/MS traces, plate reader outputs) on platforms such as Zenodo. This allows the community to verify, re-evaluate, and re-curate the data from first principles.
• Open Model Building: We build our models openly using open-source software (MIT/Apache 2.0). Trained model weights and datasets are released under CC BY 4.0 on repositories such as Hugging Face and GitHub, along with evaluation code, exact data splits, and model cards detailing performance and limitations. Anyone can freely reproduce, retrain, or extend an OpenADMET model.

As a note, blind challenges will require that a subset of the data be held back until the challenge is complete, at which point all data will be released under CC BY 4.0.

4. Expectations for Participants & Collaborators

Trust and clarity are essential for open collaboration. We set clear expectations across operational, scientific, and collaborative fronts:

• Transparency with Funding Partners: We provide clear progress of our benchmarks and the direct impact of our open science initiatives on a regular basis.
• Transparency with Our Community: We don’t hide the messy parts of data curation or experimental error. Outliers, failed assays, and model limitations will be shared openly.
• Blind Challenge Participants: To ensure absolute fairness and scientific integrity in our blind challenges, all evaluation metrics, scoring scripts, and benchmark datasets will be clearly defined and made public. Challenge frameworks, timeline milestones, and data-handling protocols are fixed and transparent to all participants from day one. We are also always available to participants to review or adjudicate any breaches of the challenge rules and to assist with any queries.

OpenADMET has several collaborators spanning various scientific disciplines. While we encourage all collaborators to participate in our vision for open science, we understand that this may not be the reality for everyone. Naturally, questions may arise for folks who wish to collaborate with OpenADMET or use our outputs.

• Independent groups using OpenADMET outputs (data, models, assays, etc.): This is evidence that OpenADMET is working! Please continue! If you have any questions or would like to discuss anything further, please don’t hesitate to reach out. While we appreciate openness among users of our data, independent groups are free to use the data under CC BY 4.0 and publish subsequent results as they see fit.
• Direct collaborations between OpenADMET-funded work and external groups: Outputs stemming from these situations should meet the above requirements. We do not participate in or support closed-source or for-profit publishing models and understand that this might limit collaborations. However, this is a trade-off we are willing to make to advance our mission. We want to make our stance clear up front. If the collaborating group chooses to publish outside our policy, OpenADMET will publish its portion first, after which the external group can publish separately, citing the OpenADMET output.

Everyone taking part in OpenADMET, including all blind challenge participants, is expected to follow this Code of Conduct, adapted from the Open Molecular Software Foundation (OMSF) Code of Conduct. It applies to our Discord, GitHub, Hugging Face, webinars, conferences, and any other spaces where OpenADMET activities take place. The original OMSF version is available at omsf.io/resources/conduct.

Introduction

This code of conduct applies to all spaces OpenADMET manages, including IRC, public and private mailing lists, issue trackers, wikis, blogs, Twitter, and any other communication channels our communities use.

We expect everyone who participates in the OpenADMET communities formally or informally, or claims any affiliation with OpenADMET, in any OpenADMET-related activities and especially when representing OpenADMET in any role to honor this code of conduct.

This code is not exhaustive or complete. It distills our common understanding of a collaborative, shared environment and goals. We expect all members of the OpenADMET communities to follow it in spirit as much as in the letter, so that it can enrich all of us and the technical communities in which we participate.

Specific Guidelines

We strive to:

• Be open. We invite anyone to participate in our community. Please use public methods of communication for project-related messages, unless discussing something sensitive. This applies to messages for help or project-related support, too; not only is a public support request much more likely to result in an answer to a question, it also makes sure that any member of the community notices and corrects any inadvertent mistakes people answering the query may make.
• Be empathetic, welcoming, friendly, and patient. We work together to resolve conflicts, assume good intentions, and do our best to act in an empathetic fashion. We may all experience some frustration from time to time, but we do not allow frustration to result in a personal attack. A community where people feel uncomfortable or threatened is not a productive one. We should be respectful when dealing with other community members as well as with people outside our community.
• Be collaborative. Other people will use our work, and we in turn depend on the work of others. When we make something for the benefit of the project, we are willing to explain to others how it works, so they can build on the work to make it even better. Any decision we make will affect users and colleagues, and we take those consequences seriously when making decisions.
• Be inquisitive. Nobody knows everything! Asking questions early avoids many problems later, so we encourage questions, although we may redirect them to the appropriate forum. Those who receive a question should be responsive and helpful, within the context of our shared goal of improving OpenADMET data, models, and software.
• Be careful in the words that we choose. Whether we are participating as professionals or volunteers, we value professionalism in all interactions, and take responsibility for our own speech. Be kind to others. Do not insult or put down other participants. Harassment and other exclusionary behaviour are not acceptable. This includes, but is not limited to:
  • Violent threats or language directed against another person.
  • Sexist, racist, or otherwise discriminatory jokes and language.
  • Posting sexually explicit or violent material.
  • Posting (or threatening to post) other people’s personally identifying information (“doxing”).
  • Sharing private content, such as emails sent privately or non-publicly, or from unlogged forums such as IRC channel history.
  • Personal insults, especially those using racist or sexist terms.
  • Unwelcome sexual attention.
  • Excessive or unnecessary profanity.
  • Repeated harassment of others. In general, if someone asks you to stop, then stop.
  • Advocating for, or encouraging, any of the above behaviour.
• Be concise. Keep in mind that, over time, hundreds or thousands of people will read what you write. Writing a short email means people can understand the conversation as efficiently as possible. Short emails should always strive to be empathetic, welcoming, friendly and patient. When a long explanation is necessary, consider adding a summary at the top of the message.

  Try to bring new ideas to a conversation so that each email adds something unique to the thread, keeping in mind that the rest of the thread still contains the other messages with arguments that have already been made.

  Try to stay on topic, especially in discussions that are already fairly long.

• Step down considerately. Members of every project come and go. When somebody leaves or disengages from the project they should tell people they are leaving and take the proper steps to ensure that others can pick up where they left off. In doing so, they should remain respectful of those who continue to participate in the project and should not misrepresent the project’s goals or achievements. Likewise, community members should respect any individual’s choice to leave the project.

Diversity Statement

OpenADMET welcomes and encourages participation by everyone. We are committed to being a community that everyone feels good about joining. Although we may not be able to satisfy everyone, we will always work to treat everyone well.

No matter how you identify yourself or how others perceive you, we welcome you. Though no list can hope to be comprehensive, we explicitly honour diversity in age, culture, ethnicity, genotype, gender identity or expression, language, national origin, neurotype, phenotype, political beliefs, profession, race, religion, sexual orientation, socioeconomic status, subculture and technical ability.

Though we welcome people fluent in all languages, OpenADMET project development takes place in English.

The Code of Conduct, above, details our standards for behaviour in the OpenADMET community. We expect participants in our community to meet these standards in all their interactions and to help others to do so.

Reporting Guidelines

While all participants should adhere to this code of conduct, we recognize that sometimes people may have a bad day, or be unaware of some of the code’s guidelines. When that happens, you may reply to them and point out this code of conduct. Such messages may be in public or in private, whatever is most appropriate. However, regardless of whether the message is public or not, it should still adhere to the relevant parts of this code of conduct; in particular, it should not be abusive or disrespectful.

Assume good faith; it is more likely that participants are unaware of their bad behaviour than that they intentionally try to degrade the quality of the discussion. Should there be difficulties in dealing with the situation, you may report your compliance issues in confidence to openadmet@omsf.io.

If the violation is in documentation or code, for example inappropriate pronoun usage or word choice within official documentation, report these privately to the project in question at their contact email addresses, and, if you have sufficient ability within the project, resolve or remove the concerning material, being mindful of the perspective of the person originally reporting the issue.

Notes

This Code defines empathy as “a vicarious participation in the emotions, ideas, or opinions of others; the ability to imagine oneself in the condition or predicament of another.” Empathetic is the adjectival form of empathy.

This statement draws on the following for content and inspiration: CouchDB Project Code of conduct, Fedora Project Code of Conduct, Speak Up! Code of Conduct, Django Code of Conduct, Debian Code of Conduct, Twitter Open Source Code of Conduct, Mozilla Code of Conduct/Draft, Python Diversity Appendix, and Python Mentors Home Page.

Credit: This Code of Conduct is the Open Molecular Software Foundation Code of Conduct, itself based on the Apache Software Foundation Code of Conduct and modified accordingly. It is adapted here under the CC BY 4.0 license, with references to OMSF (and a leftover reference to Apache) changed to OpenADMET, and reports directed to the OpenADMET team at openadmet@omsf.io.