The Strategic Plan 2021–2023 for Ersilia

The Ersilia Open Source Initiative (EOSI) is a charity organization founded in November 2020 by three scientists who believe that scientific development has a considerable impact on the progress of Low and Lower Middle Income Countries (LMIC)

The EOSI was set up with the intent to improve the limited scientific research capabilities of LMICs in the world today. It recognizes that the majority of intellectual research and literature that are currently available are focused on the scientific possibilities of High and Upper Middle Income Countries (HUMIC) thus neglecting the LMICs who are in desperate need of such expertise.

One of the ways to overcome this deficit is by improving the research capabilities of the scientific institutions in LMICs. This would enable the development of research that is targeted specifically towards their country’s unique needs.

To begin tackling this challenge, the EOSI have come up with a strategic plan focused on these two main areas:

• Drug discovery for infectious and neglected diseases.

This would enable the LMICs to research and develop their own solutions for communicable diseases which are deemed to have a low investment potential by the pharmaceutical industry.

• Implement Artificial intelligence and Machine learning (AI/ML) tools.

The use of AI/ML tools provide the ability to analyze the large amounts of data required for experimental research in record time. It is also relatively inexpensive to implement compared to full scale approaches, thus making it suitable for LMICs.

This plan was created taking into account the needs as well as the financial and collaborative opportunities of the scientific community within the Low and Lower Middle Income regions.

The goals and objectives of the plan are:

1. Facilitate the access of AI/ML technologies to all scientists.

One hindrance to implementing AI/ML methods for a majority of the research community is a lack of computational expertise. To overcome this limitation, EOSI created The Ersilia Model Hub. The Hub is a free, user-friendly repository which contains AI/ML models for drug discovery in neglected diseases.

Objectives:

• Make the Ersilia Model Hub to the public and create an interactive website for querying models.
• Include 75 researched models to the Hub yearly to increase the utilization of third party models.
• Add 50 in-house trained models derived from scientific databases and contributors to the Hub.
• Ability to precalculate and store frequent queries to provide a fast service and save computational resources.

2. Achieve sustainability for the Ersilia Model Hub.

The Hub is the key asset of Ersilia and its importance cannot be overemphasized. It not only provides access to AI/ML research tools and models, but it serves as a method for publicizing scientific research results and means of fostering collaboration within the scientific community, thus its maintenance and growth are crucial.

Objectives:

• Source for revenue from Open Source grants to complement grants received for research.
• Leverage Open Source stakeholders for support and contributions.
• Engage in crowd-sourcing campaigns periodically.
• Establish a network of donors to enable the receipt of regular grants.

3. Implement AI/ML routines in LMIC laboratories.

As well as providing free access to scientific models via the Ersilia Model Hub, the EOSI will partner with scientific institutions in LMICs to develop specific AI/ML tools and provide the necessary assistance for their adoption.

Objectives:

• Provide AI/ML training lessons for students in the medical and data science fields.
• Obtain financial resources for establishing research stations in LMICs to ensure project sustainability.
• Partner with researchers in the field to provide continued expertise and foster the development of research facilities in the countries.

4. Increase acceptance and popularity of computational research.

EOSI aims to overcome some of the misconceptions around the use of AI/ML models and boost their adoption using exhaustive documentation of each model, use case examples and strict usage guidelines.

Objectives:

• Experimental verification of AI/ML models.
• Publications in peer reviewed journals to obtain feedback from other researchers in the field.
• Open source approach to all EOSI’s projects by making all the codes and datasets publicly available, using a permissive MIT license and publishing monthly reports on all the findings.
• Participate in congresses and seminars to engage with the scientific community and increase the visibility of the AI/ML tools used by the EOSI.

5. Improve EOSI Governance and Organogram

In order to achieve the objectives listed above and become an inclusive organization, it is imperative that EOSI involves the scientific community, beneficiaries, collaborators and stakeholders in their activities and managerial processes.

Objectives:

• Incorporate a board of trustees who share the vision of EOSI, made up of willing volunteers across different disciplines.
• Foster a community of developers to provide support to the EOSI staff.
• Train and engage students to participate in short term projects geared towards EOSI’s mission.
• Recruit IT personnel to scale the operational capacity of the Ersilia Model Hub.
• Encourage the establishment of a diverse and inclusive work environment.

This strategy lays out the direction of EOSI for the next three years. It outlines a roadmap for creating sustainable organizations which could form the basis of a strong research network in LMICs.