-
Digital Pathway Curation (DPC): a pipeline able to assess the reproducibility, consensus and accuracy in biomedical search retrieval by comparing Gemini, PubMed, and Scientific Reviewers
Authors:
Flavio Lichtenstein,
Daniel Alexandre de Souza,
Carlos Eduardo Madureira Trufen,
Victor Wendel da Silva Gonçalves,
Juliana de Paula Bernardes,
Vinicius Miranda Baroni,
Carlos DeOcesano-Pereira,
Leonardo Fontoura Ormundo,
Fabio Augusto Labre de Souza,
Olga Celia Martinez Ibañez,
Nancy Starobinas,
Luciano Rodrigo Lopes,
Aparecida Maria Fontes,
Sonia Aparecida de Andrade,
Ana Marisa Chudzinski-Tavassi
Abstract:
A scientific study begins with a central question, and search engines like PubMed are the first tools for retrieving knowledge and understanding the current state of the art. Large Language Models (LLMs) have been used in research, promising acceleration and deeper results. However, besides caution, they demand rigorous validation. Assessing complex biological relationships remains challenging for…
▽ More
A scientific study begins with a central question, and search engines like PubMed are the first tools for retrieving knowledge and understanding the current state of the art. Large Language Models (LLMs) have been used in research, promising acceleration and deeper results. However, besides caution, they demand rigorous validation. Assessing complex biological relationships remains challenging for SQL-based tools and LLM models. Here, we introduce the Digital Pathway Curation (DPC) pipeline to evaluate the reproducibility and accuracy of the Gemini models against PubMed search and human expert curation. Using two omics experiments, we created a large dataset (Ensemble) based on determining pathway-disease associations. With the Ensemble dataset, we demonstrate that Gemini achieves high run-to-run reproducibility of approximately 99% and inter-model reproducibility of around 75%. Next, we calculate the crowdsourced consensus using a smaller dataset. The CSC allows us to calculate accuracies, and the Gemini multi-model consensus reached a significant accuracy of about 87%. Our findings demonstrate that LLMs are reproducible, reliable, and valuable tools for navigating complex biomedical knowledge.
△ Less
Submitted 7 May, 2025; v1 submitted 2 May, 2025;
originally announced May 2025.
-
Topological Line Defects around Graphene Nanopores for DNA Sequencing
Authors:
Jariyanee Prasongkit,
Ernane F. Martins,
Fábio A. L. de Souza,
Wanderlã L. Scopel,
Rodrigo G. Amorim,
Vittaya Amornkitbamrung,
Alexandre R. Rocha,
Ralph H. Scheicher
Abstract:
Topological line defects in graphene represent an ideal way to produce highly controlled structures with reduced dimensionality that can be used in electronic devices. In this work we propose using extended line defects in graphene to improve nucleobase selectivity in nanopore-based DNA sequencing devices. We use a combination of QM/MM and non-equilibrium Green's functions methods to investigate t…
▽ More
Topological line defects in graphene represent an ideal way to produce highly controlled structures with reduced dimensionality that can be used in electronic devices. In this work we propose using extended line defects in graphene to improve nucleobase selectivity in nanopore-based DNA sequencing devices. We use a combination of QM/MM and non-equilibrium Green's functions methods to investigate the conductance modulation, fully accounting for solvent effects. By sampling over a large number of different orientations generated from molecular dynamics simulations, we theoretically demonstrate that distinguishing between the four nucleobases using line defects in a graphene-based electronic device appears possible. The changes in conductance are associated with transport across specific molecular states near the Fermi level and their coupling to the pore. Through the application of a specifically tuned gate voltage, such a device would be able to discriminate the four types of nucleobases more reliably than that of graphene sensors without topological line defects.
△ Less
Submitted 6 December, 2017;
originally announced December 2017.
-
Interface of graphene nanopore and hexagonal boron nitride as a sensing device
Authors:
Fabio A. L. de Souza,
Wanderla L. Scopel,
Rodrigo G. Amorim,
Ralph H. Scheicher
Abstract:
The atomically-precise controlled synthesis of graphene stripes embedded in hexagonal boron nitride opens up new possibilities for the construction of nanodevices with applications in sensing. Here, we explore properties related to electronic structure and quantum transport of a graphene nanoroad embedded in hexagonal boron nitride, using a combination of density functional theory and the non-equi…
▽ More
The atomically-precise controlled synthesis of graphene stripes embedded in hexagonal boron nitride opens up new possibilities for the construction of nanodevices with applications in sensing. Here, we explore properties related to electronic structure and quantum transport of a graphene nanoroad embedded in hexagonal boron nitride, using a combination of density functional theory and the non-equilibrium Green's functions method to calculate the electric conductance. We find that the graphene nanoribbon signature is preserved in the transmission spectra and that the local current is mainly confined to the graphene domain. When a properly sized nanopore is created in the graphene part of the system, the electronic current becomes restricted to a carbon chain running along the border with hexagonal boron nitride. This circumstance could allow the hypothetical nanodevice to become highly sensitive to the electronic nature of molecules passing through the nanopore, thus opening up ways for the detection of gas molecules, amino acids, or even DNA sequences based on a measurement of the real-time conductance modulation in the graphene nanoroad.
△ Less
Submitted 24 February, 2016; v1 submitted 23 February, 2016;
originally announced February 2016.
