As the manuscript is currently under review, I welcome any feedback or question 😁

This project was a massive collaborative effort between @dukepress.bsky.social, @caltech.edu, and @ucsandiego.bsky.social. A big thank you goes to the PIs who made this possible - Rima Kaddurah-Daouk, Sarkis Mazmanian, Rob Knight, and @pieterdorrestein.bsky.social - and all the collaborators! 13/n

We believe this resource and its associated tools will aid future research in the pathophysiology of AD and in the broader context of #biomedicalresearch. Additionally, this dataset contains hundreds of yet-to-be-characterized metabolites that researchers can explore. 12/n

We then reprocessed 1.5k blood samples collected by the Alzheimer Gut Microbiome Project (AGMP) and found phenylacetyl-carnitine to be correlated with #aging, #cognitive impairment, #physical exercise and AD #biomarkers in the cerebral spinal fluid (CSF). 11/n

Looking for these previously #uncharacterized carnitines, we observed phenylacetyl-carnitine to be enriched in inflammatory diseases. This workflow can be applied to any kind of (un)annotated MS/MS spectrum researchers are interested in. 10/n

We then developed #tissueMASST to translate animal findings to human datasets. This MS/MS search tool comprises ~50k LC-MS/MS samples captured from animal models and humans with associated metadata, such as tissue type, disease status, sex, and age. 9/n

Given their possible biological relevance, we did some magic 🪄 and annotated them as benzoyl-carnitine, phenylacetyl-carnitine, and phenylpropionyl-carnitine. We #synthesized the standards and their MS/MS spectra are now part of the GNPS libraries. 8/n

These molecules were classified as modulated by the microbiome and, in an external dataset, were observed responding both to #antibiotics and #prebiotics. Additionally, they positively correlated with #inflammation markers of the mesenteric lymph nodes of the mice. 7/n

Specifically focusing on #carnitines, which were dysregulated across multiple tissues, we generated a cross-tissue molecular network to capture a holistic view. We observed 3 putative carnitines that were not annotated via GNPS MS/MS libraries. 6/n

Having both SPF and GF animals also allowed us to putatively identify if molecular alterations were driven by the #genotype, the #microbiome, or both. Information available for all the molecular features differentiating AD animals from WT. 5/n

Thousands of metabolic features were driving these separations across tissues. Alterations were observed in the #bileacids pool (obviously😉) but also in N-acyl lipids, B vitamins, neurotransmitters, polyamines, and several other classes of molecules. 4/n

Murine models of #Alzheimers Disease (AD) present distinct metabolic profiles across five tissue types captured by this atlas - brain, serum, liver, cecum, and feces. 3/n

We have generated an untargeted #metabolomics atlas from more than 2,000 samples collected from tissues of 300 mice with different genetic backgrounds (3xTg, 5xFAD, C57BL/6J), colonization conditions (SPF, GF), sex, and age. Data and metadata are publicly available. 😁 2/n