MRS Meetings and Events

 

DS03.04.08 2023 MRS Fall Meeting

A Universal Language for Experimental Protocol Reporting and Community Driven Platform for Enforcement of Field Standards

When and Where

Nov 28, 2023
4:15pm - 4:30pm

Sheraton, Second Floor, Liberty B/C

Presenter

Co-Author(s)

Robert Epps1,Robert White1,Joseph Berry1,2

National Renewable Energy Laboratory1,University of Colorado Boulder2

Abstract

Robert Epps1,Robert White1,Joseph Berry1,2

National Renewable Energy Laboratory1,University of Colorado Boulder2
Many scientific disciplines are undergoing a crisis of reproducibility, and chemistry and materials science are no exception. Despite widespread adoption of the FAIR (Findable, Accessible, Interoperable, Reusable) guidelines in 2016, there is no indication of the issue lessening, and countless hours and resources continue to be wasted on failed protocol replication. Research communication is simply not being carried out in a way that facilitates complete knowledge transfer. With the rapid growth of new and innovative automated experimental platforms, the scientific community is under the threat of even more diverse and incompatible data reporting formats that conflict with FAIR principles and are neither machine-to-machine nor machine-to-human readable. With no clear system for the communication of research procedures between and among machines and researchers, there is a critical need for a robust and flexible universal protocol standard.<br/>In this work, we present the Universal Workflow Language (UWL). UWL is a directed graph-based data entry language that enables the reporting of any arbitrary procedure in a machine-readable format. The system replicates the sequence of physical actions taken in a protocol and translates them into a data rich workflow architecture. Counter to the more traditional tabular data entry format, UWL enables flexibility to match the diverse and unique methods in which protocols are carried out in the real world. Instead of attempting to compress a full complex protocol into a series of tabulated parameters, UWL retains the initial complexity of a process.<br/>Complementary to UWL, we have developed a user interface for management and recording of protocols in the UWL format with an emphasis on minimal user inputs and maximum data density. The entry management tool allows users to create and edit UWL entries with a drag and drop block-based interface. The software allows users to import, visualize, and modify shared workflows in a more accessible format. It also includes automated error checking and validation, and it enables experimentalists to create a framework for data entry through automated/robotic experiment conduction. Additionally, the tool transcribes the workflows into both plain text and interactive tabulated formats. The plain text system reads the UWL protocol and transcribes it into a natural language protocol with corresponding material and tool details and specifications. Through the language feature, users may validate their modifications to workflows in real-time and display protocols in a variety of spoken languages.<br/>Finally, the user interface is linked to a field reporting standards system, driven by the research community. The standard system enables researchers to create and implement clear guidelines on protocol reporting standards so that procedures meet minimum detail requirements and adhere to FAIR principles. Standard creators can specify a library of workflow segments and designate error messages for missing or incomplete parameter reporting for their field of research. Every submitted standard undergoes a forum-based peer-reviewed certification process, where it is given a rating to designate comprehensiveness and field approval. This system allows researchers and manuscript reviewers to quickly evaluate whether a protocol is reported with sufficient detail without setting arbitrary constraints on the type of protocols that can be reported.<br/>UWL and the accompanying software offers comprehensive, standardized, and machine-readable data entry for both human and robotic researchers, enabling greater detail and precision in communal database generation efforts. Application of these tools would likely reduce the impact of reproducibility challenges in human led research, improve universal communication of protocols, and facilitate more effective autonomous materials research.

Keywords

standards

Symposium Organizers

James Chapman, Boston University
Victor Fung, Georgia Institute of Technology
Prashun Gorai, National Renewable Energy Laboratory
Qian Yang, University of Connecticut

Symposium Support

Bronze
Elsevier B.V.

Publishing Alliance

MRS publishes with Springer Nature