NAR 2026 manuscript: sync M02_introduction.tex with Overleaf edits

Papers/NAR_Update_2026/latex/sections/M02_introduction.tex

@@ -1,7 +1,9 @@
 \section{Introduction}
 
-The ModelSEED Biochemistry Database provides mass- and charge-balanced reaction networks reconciled across major biochemical resources for metabolic reconstructions~\cite{seaver2020}. Since its 2020 release, the database has been widely adopted for genome-scale metabolic modeling and integration. For example, GEMsembler utilizes the shared ModelSEED namespace to resolve cross-tool identifier mappings~\cite{gemsembler2025}, and Yeast9 incorporates its metabolite values to parameterize reaction Gibbs energies across the network~\cite{yeast9}.
 
-This update reports the growth of the database, and an expanded approach in handling the layer of thermodynamics data allowing us to publish every source with its own uncertainty rather than promoting one. While over $70\%$ of the reactions with multiple sources of data agree on reaction direction, any disagreement is published for review by researchers to decide which numbers and approach carry weight.
-
-Here we also describe a grading approach incorporating the uncertainties of the data across different sources, an integration of atom-mapping data to enable MFA analyses, and a structure-curation pipeline built to resolve conflict between structures from different databases. Finally, we expanded our search interface at https://modelseed.org for dissemination of the results.
+The ModelSEED Biochemistry Database provides mass- and charge-balanced reaction networks reconciled across major biochemical resources for metabolic reconstructions~\cite{seaver2020}.
+Since its 2020 release, the database has been widely adopted for curating and parameterizing genome-scale metabolic models (GEMs): e.g. GEMsembler utilizes it for cross-mapping between resources ~\cite{gemsembler2025} and Yeast9 sources Gibbs energies from it for parameterizing GEMs ~\cite{yeast9}.
+This update reports both expanding the database with more reaction biochemistry and a layered handling of thermodynamic sources and reaction directionalities that communicates the breadth of thermodynamics data, where confidence of reaction directionalities are expressed as a grade based on the strength of supporting data.
+While $>70\%$ of sources agree on reaction directionality, disagreements are displayed for investigators to interpret.
+Atom-mapping data is further integrated to enable MFA analyses, and a structure-curation pipeline is built to resolve conflicts between structures from different databases.
+Finally, we expanded our search interface at https://modelseed.org to facilitate user queries of the resource.