Skip to main content


Feb 20, 2024
# Topics

Follow Us

Unpacking insights & answering your questions

Feb 20, 2024

In March, our team of experts hosted our recurring webinar focused on the Horizon Europe Open Science requirements, which we now organise three times a year. The webinar was a real success, with 83% of the attendees rating it between 80-100% in terms of satisfaction. The overwhelming response led us to upgrade our Zoom licence to accommodate the 882 registrations!

On the day of the event, 592 individuals attended, which translates to an impressive 67% attendance rate. While a vast majority (79%) of attendees attended the whole duration of the presentation (60 minutes), we received feedback that the length of the webinar was not explicitly communicated, and some people could not attend the last half hour of Q&As. We still received 75 questions from participants and with this article, our team of experts, Jonathan England, Giulia Malaguarnera and Shanmugasundaram Venkataraman, provides, as promised, written answers to these questions.

Regarding future content, attendees expressed a strong interest in training on how to write a DMP and more concrete examples of DMPs and successful grant proposals. We are already planning to develop and incorporate both of these requests as content and/or courses on OpenPlato. Overall, we are delighted with the success of the event and the valuable feedback we received.

Get ready for our next webinar on Monday, 03 July 2023, at 12 pm CEST.

Check the Slides and recording of the Webinar that took place in March 2023.

Questions & Answers

Is there something similar to the H2020 online manual for Horizon Europe?
Is there a DMP template for Horizon Europe? / Is there a model or questionnaire to help creating the DMP?

Horizon Europe reference documents > Templates & Forms > Project reporting templates (direct download of a docx file)

You can also use a DMP creation tool to help you in the process. For instance, in Argos the template is included with guidance. You will find a short and a long version of the template. You can join the Argos community call if you need more information.

Are you planning to create a DMP catalogue with good examples?

There is no such plan yet.


The disclaimer is quite long and it is common to have little space in the acknowledment part, is it possible to reduce it?
“Funded by the European Union. Views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the European Union or [name of the granting authority]. Neither the European Union nor the granting authority can be held responsible for them.”

From a pure requirements point of view, the disclaimer must be added word-to-word to any dissemination part of the project. We would recommend talking directly with your Project Officer and/or the publisher to see if a solution can be arranged on a case-by-case basis.

Are there specific formulations to include the acronym and GA number? In H2020, this was included in the GA itself, but now no further mention of the acronym / GA number is stated.

There is currently (28/03/23) no precise sentence mentioned in the official documents. As long as the grant project name, the project number, the acronym and the mention of “Horizon Europe funding” are included, you are alright.

Please note that these must appear in the metadata (the fields describing the publication/dataset). Many ‘trusted repositories’ will give you the option to search for and select your project. If so, you do not need to add the sentence in the acknowledgement, although it is a better practice to do so for readers to easily identify the project.

Suggested sentence: “This project has received Horizon Europe funding under the [ACRONYM] grant agreement Nº [number]”
For data sharing, is it mandatory to deposit in a repository at least the data underpinning research papers published during the project (it was the case for H2020)?
You should deposit all of the data generated/collected during the project, irrespectively of whether it was used in a publication or not.

If data is linked to a publication, it should be made available at the latest on the date of publication. When uploading the data, it is important to mention the persistent identifier (usually a DOI) provided by the publisher, so links can be made between the two. If possible, you should also mention the DOI of the dataset on the publication/publisher’s website.

Please note that some publishers might require you to also deposit the underlining data on their website. Most of those do not follow the requirements of trusted repositories and are, therefore, not compliant. You would still need to upload your data separately on a trusted repository.
Is it mandatory to publish the metadata even if the associated data are not yet open? / Should the metadata be published at several instances during the project as soon as a dataset has been generated?
Yes. You must differentiate between uploading your data to make it discoverable versus making it available for others to access and reuse. In most cases, you will not (and should not) be making your data open when it is generated/collected. But you are now required to inform of its existence online through a repository, at the latest, before the end of the project. This helps with discoverability because it will be referenced earlier, could help with citation, lead to new partnerships/networking opportunities, etc. At a later stage, you will then make the data openly available for others through the repository. Note that ‘data’ includes raw data, to the extent technically feasible, but especially if it is crucial to enable reanalysis, reproducibility and/or data reuse. If you use DMP tools such as ARGOS, you can update the different versions you are producing during the project.
In Horizon Europe I cannot publish with an embargo period, right?

No. Any peer-reviewed publication must be deposited in a repository at the latest on the day of publication. Note that when publishers have two publication dates (e.g. online and per issue), the earliest date should be taken into account (usually the online release).

In the grant proposal, we cannot cite articles which are not available in Open Access. Many older papers are still not in OA and not all countries care about OA to the same extent. Is there anything we can do about this?
It will depend on the publisher, but in many cases, you will be able to make a version available in Open Access. 1. Check on SherpaRomeo whether the publisher allows you to upload the ‘Accepted Version’ on a subject and/or institutional repository. 2. Check the embargo period set by the publisher to the publication date of your work. If the embargo is over, you are free to upload it to a repository. 3. Select a repository and upload the Author Accepted Manuscript (i.e. the peer-reviewed version before being edited by the publisher). Check on SherpaRomeo and select the correct licence (e.g. CC BY-NC-ND) set by the publisher.

Some advice: * upload on the institutional repository of where you were when you published that work (some institutions might not allow you to deposit new work after you have left) * as a second choice, select a subject-specific repository relevant to your field - check OpenDOAR * otherwise, select a generic repository such as Zenodo * make sure you are uploading the Author Accepted Manuscript (i.e. peer-reviewed version) and not the Version of Record (i.e. the publisher’s final version). Note that in some cases you might be able to upload the Version of Record, but for simplicity we presented the most straighforward approach


Is there a list of trusted repositories available?
  • For publications, you can select a subject-specific repository by looking at OpenDOAR
  • For data, you can select a subject-specific repository by looking at re3data or OpenAIRE Explore
  • Alternatively, you can use a generic repository such as Zenodo
Are arXiv, biorxiv, psyXiv, etc., trusted repositories?

Those are pre-print servers where you upload your work before being peer-reviewed. The requirement is for the peer-reviewed version (i.e. Author Accepted Manuscript or Version of Record) to be made Open Access. If the pre-print server allows you to update your pre-print to the AAM/VoR once it has been published, then you would be compliant. If not, you would still need to upload it to a different repository.

What are other general repositories like Zenodo?

You can browse other repositories by looking on re3data or OpenAIRE Explore

Regarding data repositories, in Horizon Europe documentation you often find the clause that, if required in the call conditions, the repository must be federated in the EOSC in compliance with EOSC requirements. Which repositories are EOSC federated?

A list of the services offered by EOSC, including for storage and processing of research data, can be found in the EOSC Portal.

Regarding trusted repositories, institutional repositories are ok under the condition “use of PID” - what does this mean exactly? Is it required that the repository produces PIDs? Or is it sufficient that the institutional repository incorporates the PIDs of the publishers in the recorded metadata?

The publication/dataset itself must be given its own persistent identifier on the repository. This is to ensure its long-term preservation. You should still mention the PID from the publisher to create links between the two.

While DOI is nowadays the most common PID, any other type of PID is accepted (e.g. handle).

You should not just link the git repository but should download the whole code and its documentation and upload it to a trusted repository such as Zenodo. Make sure to link the git repo on Zenodo and vice-versa, providing the DOI given by Zenodo.

Intellectual Property Rights

Can we use this presentation and recording as a source to prepare something similar in our local language?

Yes. You will notice the CC-BY logo at the bottom of the presentation, which indicates you can share, reuse, and modify the content as long as you cite “England & Malaguarnera 2022. 10.5281/zenodo.7324363 under CC-BY 4.0”

If I want to share the presentation (slides) with French-speaking institutions and people, have I to translate it into French by myself or is it planned to be translated by the OpenAIRE community/coordinators?

It is not planned yet. Since the presentation is under a CC-BY 4.0 licence, you are welcome to translate it and cite “England & Malaguarnera 2022. 10.5281/zenodo.7324363 under CC-BY 4.0”

Can we use a CC BY-NC-ND licence for publications? or is it mandatory the use of a CC-BY licence?

No. It has to be licensed under a CC-BY licence. It would actually not be to your benefit as a researcher to share your work under a more restrictive licence. The only exceptions are monographs or other longer formats.

For the longer formats, are CC BY-NC-ND licences allowed (not just NC or ND, but both)?

Yes. You could select CC BY-NC, CC BY-ND or CC BY-NC-ND.

CC BY also requires to indicate the changes that were made in the reused document, doesn´t it?

Yes. You must give appropriate credit, provide a link to the licence, and indicate if changes were made.

Is there any way to prove that the publication uploaded in the repository is the Author Accepted Manuscript version (previously called postprint)?

To the best of our knowledge, there is no tool to check the version uploaded. There are key elements which can indicate it is not the Version of Record and not a preprint, but even for someone trained or experienced; it is not always possible.

If data is shared under a CC0 licence, will someone using it have to cite that they have used it?

No. While it is best practice to do so, it is not legally required. However, especially in research where it is already common practice to cite others’ work, a CC0 licence does not mean you cannot “request attribution in accordance with community or professional norms and standards” (see CC0 FAQ).

There is no specific licence to use for software, although it is suggested to use an equivalent to a CC0 licence. They strongly recommend “those listed as free by the Free Software Foundation and listed as open source by the Open Source Initiative” Using a licence selector tool might help choosing which licence to apply.

Domain or project-specific

In many cases, the data has commercial value. How do you define this?

Commercial value falls under the ‘Exploitation of results’ section of the grant. You must “take measures aiming to ensure exploitation of your results, either by yourself (e.g. a beneficiary owning results uses them directly) or indirectly by others (other beneficiaries or third parties, e.g. through licencing or by transferring the ownership of results).” (AGA v0.2 - Annex 5)

“Exploitation (as defined) means the use of results in further research and innovation activities other than those covered by the action concerned, including, among other things, commercial exploitation such as developing, creating, manufacturing and marketing a product or process, creating and providing a service, or in standardisation activities.” (AGA v0.2 - Annex 5).
If my Horizon Europe project has a market deployment focus, the Open Science requirements are not applicable?

If the project generates/collects data, you have to follow the Open Science requirements. The main difference is that you might not make the data openly available (or set an embargo period). But you would still need to upload your data to a trusted repository to make it discoverable. This must be done as soon as possible and, at the latest, before the end of the project. In exceptional cases in which specific constraints apply (e.g. security rules), deposition can be delayed beyond the end of the project. You will need to clearly explain why you consider it necessary to close the data (remember that the default is open) in the Data Management Plan.

Having said that, the European Commission is very clear that, in some cases, it may be advisable to protect an invention by keeping it confidential as a trade secret or to postpone the filing of a patent (or other IPR) application.

Any scientific publications will also have to comply with the Open Science requirements. You might be required to “provide (digital or physical) access to data or other results needed for the validation of the conclusions of these publications, to the extent that your legitimate interests are safeguarded and constraints are taken into account”(e.g. agreements with relevant confidentiality).

Note that results generated by associated partners still need to respect all of the Open Science obligations. Also, if one year after the end of the action, key exploitable results are not exploited, you will need to use the Horizon Results Platform.

In case of public emergencies, “beneficiaries must grant non-exclusive licences — under fair and reasonable conditions — to legal entities that need the research output to address the public emergency and commit to rapidly and broadly exploit the resulting products and services on fair and reasonable conditions. This obligation will apply for a period of time specified in the request and up to four years after the end of the action. The duration of this obligation and the fairness and reasonability of the licences will be assessed on a case-by-case basis and will depend on the specific circumstances of the public emergency, the context of each project and the nature of its results.” (AGA v0.2 - Annex 5.2.3)

Are there any examples of how these principles can be implemented in a Horizon Europe proposal in a project for product development?

For questions relating to the development of grant proposals, we suggest contacting your National Contact Point.

Could you please give examples of data in a philosophy project?
Research data is all of the information that you use as an integral part of your research and used to validate your research findings.
These can take many forms, for instance: text documents, scanned manuscripts, digital audiotapes/videotapes, scanned photographs or films, etc.
Are excluded: data produced by non-research activities (e.g. teaching, administrative tasks, mailbox backup, timesheets).
Rights Retention Strategy

About Green Open Access: some universities propose using a Rights Retention Strategy in which authors ensure that the Author Accepted Manuscript (AAM) is under a CC-BY licence by adding a ‘rights retention statement’ to the submitted manuscript. If the journal accepts, authors can immediately upload it to any repository.

According to the Horizon Europe Grant Agreement, immediate open access needs to be provided to all publications. In this framework, is there room for the use of Green Open Access?

Yes. As long as you retain copyright on your Author Accepted Manuscript (i.e. the peer-reviewed version before being edited by the publisher) and apply a CC-BY 4.0 licence, you can share and upload that version without any restrictions wherever you want, including a trusted repository. Note that in this case, it does not matter if another licence is applied to the Version of Record.

You can use this route by adding a rights retention statement on your manuscript upon submission to the publisher. Not all publishers will accept it and you want to make sure this is agreed upon before the peer-reviewing process. Here is a slightly modified version of the statement the European Commission suggests adding:

“This work was funded by the Εuropean Union under the Horizon Europe grant [project acronym][grant number]. As set out in the Grant Agreement, beneficiaries must ensure that at the latest at the time of publication, Open Access is provided via a trusted repository to the published version or the final peer-reviewed manuscript accepted for publication under an open licence. For this purpose, the author has applied a [CC-BY 4.0]/[CC BY-NC/CC BY-ND/CC BY-NC-ND 4.0 (as it is a long-text format)] licence to any Author Accepted Manuscript (AAM) version arising from this submission.”

Note that this ‘Green route’ approach for Open Access is only possible for subscription-based and hybrid journals. In full-Open Access publishing venues, you will have to pay for the Article/Book Processing Charges (if any).

Do you think journals will accept this Rights Retention Strategy? / Do you know any cases in which the Rights Retention Strategy has been implemented?

“We only know of two journals that reject articles that are using the Rights Retention statement: Blood, and Blood Advances” (statement from cOAlition S, 21/03/2023).

Publishers use various strategies to prevent you from using the rights retention statement: * during the submission process; some publishers will explicitly say you cannot submit your work if a rights retention statement has been added to the manuscript; * others have adapted the submission workflow so that if a rights retention statement is present on your manuscript, you will be required to go the payment route (i.e. Article Processing Charges). Remember that APCs for hybrid journals cannot be claimed under your grant. By accepting this ‘contractual agreement’ to pay APCs during the submission process, you will have to find another way to fund these costs if your work gets accepted.

Is there any approval we need from the Project Office to proceed with the rights retention statement?

No. You may add the statement to any work you submit for publication.

Open Research Europe

In Open Research Europe, can the automatic deposit in Zenodo be de-activated in the case that our institution asks us to deposit in the institutional repository?
The deposit on Zenodo does not prevent you from also uploading it on your institutional repository. It might actually event boost its discoverability. Remember to always mention the persistent identifier provided by the publisher so connections can be made with your institutional repository.

Note that “Open Research Europe plans to syndicate published content with institutional and national repositories in Europe and will gradually add to the options of repositories that we send content to, following the preference of the author.” (Open Research Europe FAQ, retrieved 28/03/2023).

It was mentioned that Open Science Europe is indexed in Scopus. Is it also indexed in the Web of Science?

It is not yet the case (as of 28/03/2023), but it is planned.

What is the uptake of Open Research Europe? Is it used a lot?

“In the past two years, over 1,800 authors from 340 different institutions and 45 different countries have published their research with Open Research Europe, with over 340 articles covering everything from battery performance for electric vehicles to Renaissance shipbuilding. […] Since its launch, research published on Open Research Europe has amassed over 28,000 article views, over 7,000 article downloads, over 400 article citations. Some of these citations include coverage in news outlets such as Forbes and the Columbia Journalism Review, and in policy documents of the Food and Agriculture Organization of the United Nations and the International Renewable Energy Agency.”(Celebrating two years of Open Research Europe, Mckenny 2023).

Can you say something about the average time from submission to publication in Open Research Europe?

The average time should be of about 10 days.

ORE papers are starting to be indexed in Scopus but it takes a very long for them to show up. Is this going to remain taking so long, or is it a startup issue?

We do not think it takes much more time than any other indexing, but we will check with ORE and update this answer.

Is it possible to become a reviewer for submitted papers in Open Research Europe?

“Reviewers are formally invited by the Open Research Europe editorial team (as is the case with most journals). They are selected based on suggestions from the authors, and their suitability as a reviewer for the article is checked before they are invited to contribute a review. Prospective reviewers who have collaborated or are currently collaborating with any of the authors are not eligible to review the article in question. Reviewers are also asked to declare any competing interests.” (Open Research Europe FAQ, retrieved 28/03/2023)

Is there any expectation that the Open Research Europe metrics will be accepted by national evaluation bodies instead of/along with the usual ones, such as JCR…? / Ultimately, the success or failure of Open Research Europe will depend on whether they are considered “high impact”. Will young researchers’ CVs showing Open Research Europe submissions be valued or triaged during the selection process? Will the EU/can the EU influence this?

“Articles published on this platform are associated with a variety of quantitative and qualitative metrics to provide open, article-level information, allowing the article to be evaluated on its own merit. Transparent reviewing of all articles through open peer review also provides an important assessment at the article level, in addition to more traditional measures such as views, downloads and citations. The scope of Open Research Europe is to encourage the publication of high-quality research from Horizon 2020, Horizon Europe and Euratom funding. The Commission expects funders and institutions to recognize the platform as a comparable publication venue to traditional journals.” (Open Research Europe FAQ, retrieved 28/03/2023)

Sharing research data

Do you know if editorials and journals are imposing requirements on the open access of research data needed to validate a publication as they do with the open access of publications?
There are currently no standards, but many publishers are adding guidelines for authors. You should check this every time you want to publish. Please note that when publishers ask you to deposit the underlining data on their website, most of those do not follow the requirements of trusted repositories and are therefore not compliant. You would still need to upload your data separately on a trusted repository.
This is not well defined, “open as possible and close at necessary”
Each discipline and each research project vary greatly in terms of data management. While the default should be to open the data, you can decide some parts of the project might need to be permanently or temporarily closed. This must be explicitly justified, explaining in the DMP the exception(s) under which you choose to or must restrict access to some or all of the research data.

The accepted exceptions are: * if providing open access is against the beneficiary’s legitimate interests, including regarding commercial exploitation; * if it is contrary to any other constraints, such as data protection rules, privacy, confidentiality, trade secrets, Union competitive interests, security rules, intellectual property rights or would be against other obligations under the Grant Agreement.

Who decides whether the decision to not provide open access to data is justified? The project officer when you submit the DMP?
During the submission process, the reviewers will be evaluating the Open Science aspects of your proposal. Plans on not opening some or all of the research data need to be explained in detail. You are always on the safer side if you decide to at least publish some research outputs (which can also mean source code of analysis scripts, research material, developed concepts, guides, etc.). During and at the end of the project, it will be the Project Officer that will evaluate your DMP and, therefore, the reasoning behind not opening some or all of the data.

Data Management Plan

To what extent is the DMP about data management during research or about data sharing after research?

The DMP should cover both the internal data management during the project and the sharing of the data. For instance, you might not use the same file format during the project while exchanging with partners (e.g. sharing a spreadsheet as an xlsx file) and once you are formatting the data to upload on a repository (e.g. spreadsheet data as a CSV file). It might be easier at the start when planning the DMP to differentiate between the two.

I am supporting numerous EU projects for Data management and DMP and also doing DMP training, but I rarely see real reviewing of DMP by EU experts. Are the DMPs really reviewed? / Do you have a view on the review process of DMPs by the project officers (or externals)? How many initial DMPs (after 6 months) are accepted/rejected? What are the reasons for rejection?

Yes. DMPs are reviewed by the Project Officer and/or contracted independent experts. Whether or not it is rejected will depend on how thorough you are in the DMP, the project itself (e.g. privacy issues, closing the data for commercial purposes, etc. may require further discussions) and the reviewer’s evaluation style (e.g. supportive, controlling, directive).

You may be interested in an article which analysed DMPs in H2020 - Spichtinger 2021.

Are there resources on how to proceed and what not to miss when updating DMPs?
New versions of the DMP should be created whenever important changes to the project occur due to the inclusion of new data sets, changes in consortium policies or external factors. The DMP should be updated at a minimum in time with the periodic evaluation/assessment of the project. If there are no other periodic reviews foreseen within the grant agreement, then such an update needs to be made in time for the final review at the latest. Furthermore, the consortium can define a timetable for review in the DMP itself.
What and where are the criteria for reviewing the DMPs? Are those from Science Europe?

The DMP is evaluated based on the same guiding questions presented in the Horizon Europe DMP template.

Is there a delay in reviewing DMPs?

As with other reporting obligations, the Project Officer will inform you of the timeline to update your DMP based on the comments made.

Are you aware of any budget impact if a bad DMP is submitted?

As with any of your other obligations, consequences of non-compliance are presented in Chapter 5 of the Model Grant Agreement. These include grant reduction, suspension and termination, administrative sanctions, etc. However, as long as you take into consideration any comments made by the Project Officer and/or reviewers and update the DMP accordingly, there is no reason any sanctions will be applied.

Can you describe the difference between different DMP deliverables? We just came across a case where after the first DMP (M6) an additional “Ethics-DMP” (M12) was required. It is just a revised version with more detailed information on the handling of sensitive/personal data?

We are not aware of any specific “Ethics-DMP”. Handling sensitive and personal data should definitely be covered in the first DMP (M6). It might be specific to your project, so you should check directly with the Project Officer.

I’m not quite sure I understand what you mean by “don’t let the project officer dig for information”. Could you expand on that?

You do not want to remain vague and general when completing the DMP. If the Project Officer does not find all the relevant information in your DMP, you will be required to revise it. Be as precise and detailed as possible.

Are DMP tools interoperable, i.e. is it possible to import freely written DMPs or written through other DMP tools? Or at least is Argos interoperable?

Not yet.

Are the new requirements for DMP reflected in Argos?

Argos is continuously being improved and updated to match the current requirements set in Horizon Europe.

In your opinion/experience, what is the most difficult part of meeting the RDM mandates?

Research Data Management has many elements to consider and it is easy to forget or not be aware of some parts of it. Without proper training (e.g. workshop on RDM and DMPs) and the help of a data management specialist (e.g. data steward, data curator, research data librarian), you might lose more time than necessary on coming up with a ‘good’ plan for your research data.

It is also very important to not forget to include legal, technology transfer and IT experts from your institution at an early stage of the project. For instance: * Data Protection Officer (DPO) - consulted for compliance managing personal data (e.g. GDPR); * Chief Information Security Officer (CISO) - consulted for data security and best practices for data storage and sharing; * IPR expert - consulted for copyright ownership, transfers or any other doubts relating to intellectual property rights; * Technology Transfer Officer - consulted for any commercial applications of your project outputs.

Best practices

APCs are not eligible for reimbursement for hybrid journals. Does this mean the EU prefers for authors to publish in a full OA journal with a low Impact Factor rather than a hybrid journal with a high Impact Factor?

The European Commission leaves authors the choice of publishing in any publishing venue. A journal’s metrics (e.g. Impact Factor) are not considered anymore for the proposal or evaluation of the impact of a project. The focus is now on the quality of the publication itself, including article-level metrics. When considering a publishing venue, you should therefore think about where it might get the most visibility from the community you are trying to reach.

How will the quality of the cited publications be determined by the evaluators during the proposal evaluation?

(see the previous answer)

If the impact factor of the journal is not going to be taken into account anymore, what other bibliometric indicators are going to be evaluated?

(see the previous answer)

Are there good examples of how to use standards and vocabularies when using a catch-all repository (e.g. Zenodo, with its basic metadata)? Is it possible only through an attached file (documentation)? In this case, it would not be interoperable and machine friendly.

It highly depends on the disciplines/types of data. It is important to differentiate between the two types: metadata of the dataset as a whole and metadata of the data within the dataset. 1) metadata of the dataset as a whole: * Describe the dataset on the repository as thoroughly as possible (e.g. use appropriate tags etc.), link the dataset with related publications and vice versa, and link the dataset to all relevant sub-communities if you are using Zenodo; * prefer discipline-specific repositories with more elaborated metadata if available. 2) the metadata of the data within the dataset: * provide well-documented codesheets in an open format if possible/relevant, provide xml/json files with rich metadata (linked with existing vocabularies if available) along with your dataset; * it is quite easy to use such files for machine-action of the data for further reuse of the data; * if there are any curated databases in your research domain where you could insert the metadata (either type 1 and/or type 2) then it would be a better practice to do so.


Are open access publishing costs (APCs) eligible for reimbursement?
If the publishing venue is full-Open Access, Article/Book Processing charges are eligible for reimbursement. Remember to consider these costs when drafting the budget. If you publish in a hybrid journal, Open Access costs are NOT eligible for reimbursement. You will have to find another way of covering these costs.
Note that in the context of Horizon Europe, so-called mirror and sister journals are considered Open Access publishing venues and not hybrid journals. “Mirror and sister journals are more recently established open-access versions of existing subscription journals. They may share the same editorial board as the original journal and usually have (at least initially) the same or very similar aims, scope and peer review processes and policies. These journals often have a name similar to the subscription title but a different ISSN.” (AGA v0.2 - Annex 5.2.1)
Are other publication costs (e.g. color charges, fees for overlength etc.) eligible for reimbursement?

No. You will have to find another way of covering these costs.

About the budget for data curations costs, can I include human resources (e.g. a Data Steward) for all the project or does it only cover equipment/maintenance/storage costs?
Yes. Any type of costs related to research output management are eligible as long as the eligibility conditions are fulfilled. This includes human time, external services, storage space, equipment, etc. Note that costs for drawing up the plan for the exploitation and dissemination of the results are normally NOT eligible since they will have been incurred before the start of the action to prepare the proposal. Costs that occur when revising or implementing this plan may be eligible if the eligibility conditions are fulfilled.