June 30, 2026​

​

​The ETC is seeking a practical, scalable, and scientifically robust computational tool for the prediction of solubility in multicomponent crystal systems, particularly salts and cocrystals, with sufficient predictive capability to reduce reliance on extensive experimental screening and support more efficient pharmaceutical discovery and development.​

 

The scope of work listed in the RFI highlights the open questions identified by the Solubility Modeling Working Group in the area of in-silico solubility prediction of multicomponent crystals. Interested parties are welcome to respond to any number of the sections below, especially if their expertise can only address certain sections or a portion of a given section.

​​

Download the Request for Proposal and submit your response

RFI ISSUED:  June 30, 2026

QUESTIONS on RFI DUE to ETC:  July 21, 2026

RFI RESPONSES DUE to ETC: August 21, 2026

​

Q&A - Note see FAQ document for answers to common questions​​​ (updated 8/12/2026)

• How much funding is available for this project?  There is currently no funding level determined for the project. ETC’s process is to solicit proposals and get estimates of project cost.  The ETC Team will select review and score the proposals, inviting the respondents to interview with ETC members.  Once a proposal is selected, the ETC team works with the respondent to create a SOW which defines the scope of the project, deliverables, timelines, and cost. The SOW is presented to the Board of Directors and if enough ETC members commit funding to the project it will go forward.  Please note that ETC will allow indirect costs of up to 10% of total project costs associated with personnel, staff, and supplies. All other budget categories—including equipment—are not eligible for indirect costs.

• Will ETC members provide data for this project?  Can ETC provide, or facilitate access to, an initial benchmark dataset of multicomponent (salt/co-crystal) solubility measurements with documented solid-state characterization? What volume and quality of data can realistically be expected? Is there an existing internal or consortium dataset that could be contributed under an appropriate data-sharing agreement, including anonymized or aggregated datasets where applicable? ETC companies have some internal data that is likely sufficient for benchmarking; however, we expect the collaborator to source the data required for the development of the tool.

• For the initial project, which use case should be treated as the primary acceptance use case: solvent ranking for a known crystal form, absolute solubility prediction, or in silico counter-ion/co-former screening? Can ETC identify the minimum viable scope versus future product capabilities? Minimum requirement for the project is the capability to do solvent ranking for a particular crystal form. Desired capability is the prediction of absolute solubilities for different crystal forms to be able to do in-silico counter-ion and co-crystal screening 

• Is there a target set of "must-support" solvents beyond the general description of common pharmaceutical process-development solvents, such as a predefined list from a member-company reference standard or SOP? While the exact list will be determined during scope of work (SOW) activities, ICH class 3 solvents as well as some ICH 2 solvents like MeCN, DCM, MeOH, THF, Toluene are expected to be supported. Water should also be supported, while biorelevant media is optional and preferred.  

• Is the expected project deliverable primarily a validated modeling prototype, a production-ready package, or a fully deployed enterprise application with user interface, APIs, database connectivity, authentication, deployment, and support capabilities? The expected deliverable is an application with user interface, APIs, back end, database connectivity, authentication, deployment, and support capabilities

• How many member companies are expected to participate in pilot testing and validation activities, and over what timeframe are these activities anticipated to occur? Decision on company participation in the project will be determined after the collaborator is selected and a SOW is developed.  You can view the companies interested in the project Section 2.1 of the RFI.  As for the timeframe, timing depends on when the project is scoped and brought to the ETC Board of Directors during one of its two funding cycles, one in the Fall and one in the Spring.  If approved, a project could start in Summer or Winter.

• Are there specific software platforms, tools, or technology stacks that ETC expects collaborators to leverage, or should respondents assume a technology-agnostic approach when proposing the solution architecture? Technology-agnostic approach 

• Are there specific target LIMS, ELN, MES, ERP, or other enterprise systems with which the OPC UA and API layers must interoperate, or should the integration strategy remain vendor-agnostic? Vendor-agnostic approach 

• What deployment model is anticipated for the solution, such as on-premises deployment, member-company private cloud environments, a shared multi-tenant SaaS platform, or a vendor-hosted solution? On-premises deployment and/or member-company private cloud environments .

• Is ETC expecting a single vendor to deliver the complete solution, or is the intent to assemble capabilities from multiple partners (for example, separate providers for modeling, software engineering, and systems integration)? There is no expectation from ETC on how a complete solution is delivered.  It is up to the RFI respondent to propose an approach for ETC to consider.

• What are ETC's expectations regarding ongoing model maintenance, retraining, and continuous improvement as new experimental data becomes available? Additionally, what governance framework is envisioned for contributed data, model updates, and version control? ETC's overall goal for every project is for the solution to be commercially available and supported. 

• Does ETC have a target budget range, desired project duration, or anticipated phase structure that respondents should consider when developing their proposal and cost estimate? Please review our general FAQ for answers to these questions.

• How will ownership, licensing, and usage rights be managed for background intellectual property, newly developed algorithms, source code, trained model artifacts, ETC-provided data, derived datasets, benchmark results, and consortium-specific model configurations? Please review our general FAQ for answers to these questions.

• May we assume that the 3D crystal structure of the target inference compound is known? Specifically, can we utilize structural data from Crystallographic Information Files (CIF) for model training, featurization, and solid-state interaction mapping? Minimum input requirement should be the SMILES string and stoichiometry for the multicomponent crystal. The CIF files will likely not be available during the stage of development where in-silico solubility prediction is needed.

• Given the known scarcity of high-quality solubility datasets for multicomponent systems, especially concerning diverse co-formers, can a defined portion of the project scope be dedicated to generating a relevant baseline dataset? We hope to systematically bridge this data gap using our agentic AI driven data curation pipelines alongside high-throughput experiment driven data generation. The workgroup is on board with a portion of the project scope being dedicated to dataset generation, provided that the project cost and time estimate is broken down into dataset generation and model development components

• Multi-component solutes with the same chemical composition can crystallize into different forms of crystals [eg, ranitidine hydrochloride (Zantac) has two salts forms: Form I and Form 2], and these different forms of multi-component crystals with the same chemical composition have different solubilities. Will the cif files for the multi-component crystal structures be provided as optional inputs for the solubility calculations? Minimum input requirement should be the SMILES string and stoichiometry for the multicomponent crystal. The CIF files will likely not be available during the stage of development where in-silico solubility prediction is needed