Search

Abstract There has been a growing effort to replace manual extraction of data from research papers with automated data extraction based on natural language processing, language models, and recently, large language models (LLMs). Although these methods enable efficient extraction of data from large sets of research papers, they require a significant amount of up-front effort, expertise, and coding. In this work, we propose the ChatExtract method that can fully automate very accurate data extraction with minimal initial effort and background, using an advanced conversational LLM. ChatExtract consists of a set of engineered prompts applied to a conversational LLM that both identify sentences with data, extract that data, and assure the data’s correctness through a series of follow-up questions. These follow-up questions largely overcome known issues with LLMs providing factually inaccurate responses. ChatExtract can be applied with any conversational LLMs and yields very high quality data extraction. In tests on materials data, we find precision and recall both close to 90% from the best conversational LLMs, like GPT-4. We demonstrate that the exceptional performance is enabled by the information retention in a conversational model combined with purposeful redundancy and introducing uncertainty through follow-up prompts. These results suggest that approaches similar to ChatExtract, due to their simplicity, transferability, and accuracy are likely to become powerful tools for data extraction in the near future. Finally, databases for critical cooling rates of metallic glasses and yield strengths of high entropy alloys are developed using ChatExtract.

Abstract Osteoclasts are the primary target for osteoporosis drug development. Recent animal studies revealed the crucial roles of osteoblasts in regulating osteoclastogenesis and the longer lifespans of osteoclasts than previously thought with fission and recycling. However, existing culture platforms are limited to replicating these newly identified cellular processes. We report a demineralized bone paper (DBP)-based osteoblast culture and osteoclast assay platform that replicates osteoclast fusion, fission, resorption, and apoptosis with high fidelity and analytical power. An osteoid-inspired DBP supports rapid and structural mineral deposition by osteoblasts. Coculture osteoblasts and bone marrow monocytes under biochemical stimulation recapitulate osteoclast differentiation and function. The DBP-based bone model allows longitudinal quantitative fluorescent monitoring of osteoclast responses to bisphosphonate drug, substantiating significantly reducing their number and lifespan. Finally, we demonstrate the feasibility of humanizing the bone model. The DBP-based osteo assay platforms are expected to advance bone remodeling-targeting drug development with improved prediction of clinical outcomes.

Abstract Fluorescence analysis is a fast and sensitive method, and has great potential application in trace detection of environmental toxins. However, many important environmental toxins are non-fluorescent substances, and it is still a challenge to construct a fluorescence detection method for non-fluorescent substances. Here, by means of charge transfer effect and smart molecular imprinting technology, we report a sensitive indirect fluorescent sensing mechanism (IFSM) and microcystin (MC-RR) is selected as a model target. A molecular imprinted thin film is immobilized on the surface of zinc ferrite nanoparticles (ZnFe2O4 NPs) by using arginine, a dummy fragment of MC-RR. By implementation of IFSM on the paper-based microfluidic chip, a versatile platform for the quantitative assay of MC-RR is developed at trace level (the limit of detection of 0.43 μg/L and time of 20 min) in real water samples without any pretreatment. Importantly, the proposed IFSM can be easily modified and extended for the wide variety of species which lack direct interaction with the fluorescent substrate. This work offers the potential possibility to meet the requirements for the on-site analysis and may explore potential applications of molecularly imprinted fluorescent sensors.

There has been a growing effort to replace manual extraction of data from research papers with automated data extraction based on natural language processing, language models, and recently, large language models (LLMs). Although these methods enable efficient extraction of data from large sets of research papers, they require a significant amount of up-front effort, expertise, and coding. In this work, we propose the ChatExtract method that can fully automate very accurate data extraction with minimal initial effort and background, using an advanced conversational LLM. ChatExtract consists of a set of engineered prompts applied to a conversational LLM that both identify sentences with data, extract that data, and assure the data's correctness through a series of follow-up questions. These follow-up questions largely overcome known issues with LLMs providing factually inaccurate responses. ChatExtract can be applied with any conversational LLMs and yields very high quality data extraction. In tests on materials data, we find precision and recall both close to 90% from the best conversational LLMs, like GPT-4. We demonstrate that the exceptional performance is enabled by the information retention in a conversational model combined with purposeful redundancy and introducing uncertainty through follow-up prompts. These results suggest that approaches similar to ChatExtract, due to their simplicity, transferability, and accuracy are likely to become powerful tools for data extraction in the near future. Finally, databases for critical cooling rates of metallic glasses and yield strengths of high entropy alloys are developed using ChatExtract., (© 2024. The Author(s).)