Machine Learning Predicts Biogeochemistry from Microbial Community Structure in a Complex Model System

Dutta, Avishek; Goldman, Thomas; Keating, Jeffrey; Burke, Ellen; Williamson, Nicole; Dirmeier, Reinhard; Bowman, Jeff S.

doi:10.1128/spectrum.01909-21

Cited by 17 publications

(4 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These techniques are well suited to complex, high dimensional, community structure data and can be used to extract patterns of succession and biogeochemical signatures from sequence information (Bowman, 2021). For example, the Random Forest (RF) regression model is effective at predicting biogeochemical signatures from amplicon sequence data, providing the potential for extending the data coverage of less frequently sampled key biogeochemical variables (Dutta et al, 2022). Additionally, potential microbial drivers for these processes can be identified by applying permutation to the RF models to assess the contributions of specific community members to model performance (DiMucci et al, 2018).…”

Section: Ecological Modelingmentioning

confidence: 99%

Overview of the MOSAiC expedition: Ecosystem

Fong,

Hoppe,

Team

et al. 2023

Preprint

View full text Add to dashboard Cite

An international and interdisciplinary sea ice drift expedition, the ‘The Multidisciplinary drifting Observatory for the Study of Arctic Climate‘ (MOSAiC), was conducted from October 2019 to September 2020. The aim of MOSAiC was to study the interconnected physical, chemical and biological characteristics and processes from the atmosphere to the deep sea of the central Arctic system. The ecosystem team addressed current knowledge gaps and explored unknown biological properties over a complete seasonal cycle focusing on three major research areas: biodiversity, biogeochemical cycles and linkages to the environment. In addition to the coverage of core properties along a complete seasonal cycle, dedicated projects covered specific processes and habitats, or organisms on higher taxonomic or temporal resolution. A wide range of sampling approaches from sampling, sea ice coring, lead sampling to CTD rosette-based water sampling, plankton nets, ROVs and acoustic buoys was applied to address the science objectives. Further, a wide range of process-related measurements to address e.g. productivity patterns, seasonal migrations and diversity shifts were conducted both in situ and onboard RV Polarstern. This paper provides a detailed overview of the sampling approaches used to address the three main science objectives. It highlights the core sampling program and provides examples of two habitat- or process-specific projects. First results presented include high biological activities in winter time and the discovery of biological hotspots in underexplored habitats. The unique interconnectivity of the coordinated sampling efforts also revealed insights into cross-disciplinary interactions like the impact of biota on Arctic cloud formation. This overview further presents both lessons learned from conducting such a demanding field campaign and an outlook on spin-off projects to be conducted over the next years.

show abstract

Section: Ecological Modelingmentioning

confidence: 99%

Overview of the MOSAiC expedition: Ecosystem

Fong,

Hoppe,

Team

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Tree-based algorithms, which involve various conformations and combinations of decision trees, have been shown to perform well on available microbial community datasets. These studies have ranged from predicting strength and direction in pairwise interactions of bacteria (DiMucci et al 2018 ), prediction of hydrogen sulphide production (Dutta et al 2022 ) and dissolved organic carbon levels (Thompson et al 2019 ) based on microbial community structure as an input, prediction of soil microbial community structure from environmental parameters (Peng et al 2022 ), and ranking of phenotypes important to sugar consumption in wine yeast communities (Ruiz et al 2023 ). It is therefore clear that these algorithms may aid in the quest to predicting temporal dynamics in microbial communities.…”

Section: Introductionmentioning

confidence: 99%

Gradient boosted regression as a tool to reveal key drivers of temporal dynamics in a synthetic yeast community

Conacher,

Watson,

Bauer

2024

FEMS Microbiology Ecology

View full text Add to dashboard Cite

Microbial communities are vital to our lives, yet their ecological functioning and dynamics remain poorly understood. This understanding is crucial for assessing threats to these systems and leveraging their biotechnological applications. Given that temporal dynamics are linked to community functioning, this study investigated the drivers of community succession in the wine yeast community. We experimentally generated population dynamics data and used it to create an interpretable model with a gradient-boosted regression tree approach. The model was trained on temporal data of viable species populations in various combinations, including pairs, triplets, and quadruplets, and was evaluated for predictive accuracy and input feature importance. Key findings revealed that the inoculation dosage of non-Saccharomyces species significantly influences their performance in mixed cultures, while Saccharomyces cerevisiae consistently dominates regardless of initial abundance. Additionally, we observed multispecies interactions where the dynamics of Wickerhamomyces anomalus were influenced by Torulaspora delbrueckii in pairwise cultures, but this interaction was altered by the inclusion of S. cerevisiae. This study provides insights into yeast community succession and offers valuable machine learning-based analysis techniques applicable to other microbial communities, opening new avenues for harnessing microbial communities.

show abstract

“…For example, MLs can recognize patterns better representing the individual risk compared to classical surgical risk scores ( 6 ). ML includes various types, such as support vector machine (SVM) ( 7 , 8 ), random forest (RF) ( 9 ), decision tree (DT) ( 10 – 12 ), and so on. Different ML has its specialty and shortcoming.…”

Section: Introductionmentioning

confidence: 99%

Multiple-model machine learning identifies potential functional genes in dilated cardiomyopathy

Zhang

Lin²,

Wang³

et al. 2023

Front. Cardiovasc. Med.

View full text Add to dashboard Cite

IntroductionMachine learning (ML) has gained intensive popularity in various fields, such as disease diagnosis in healthcare. However, it has limitation for single algorithm to explore the diagnosing value of dilated cardiomyopathy (DCM). We aim to develop a novel overall normalized sum weight of multiple-model MLs to assess the diagnosing value in DCM.MethodsGene expression data were selected from previously published databases (six sets of eligible microarrays, 386 samples) with eligible criteria. Two sets of microarrays were used as training; the others were studied in the testing sets (ratio 5:1). Totally, we identified 20 differently expressed genes (DEGs) between DCM and control individuals (7 upregulated and 13 down-regulated).ResultsWe developed six classification ML methods to identify potential candidate genes based on their overall weights. Three genes, serine proteinase inhibitor A3 (SERPINA3), frizzled-related proteins (FRPs) 3 (FRZB), and ficolin 3 (FCN3) were finally identified as the receiver operating characteristic (ROC). Interestingly, we found all three genes correlated considerably with plasma cells. Importantly, not only in training sets but also testing sets, the areas under the curve (AUCs) for SERPINA3, FRZB, and FCN3 were greater than 0.88. The ROC of SERPINA3 was significantly high (0.940 in training and 0.918 in testing sets), indicating it is a potentially functional gene in DCM. Especially, the plasma levels in DCM patients of SERPINA3, FCN, and FRZB were significant compared with healthy control.DiscussionSERPINA3, FRZB, and FCN3 might be potential diagnosis targets for DCM, Further verification work could be implemented.

show abstract

Machine Learning Predicts Biogeochemistry from Microbial Community Structure in a Complex Model System

Abstract: Microbial communities control many biogeochemical processes. Many of these processes are impractical or expensive to measure directly.

Cited by 17 publications

References 45 publications

Overview of the MOSAiC expedition: Ecosystem

Overview of the MOSAiC expedition: Ecosystem

Gradient boosted regression as a tool to reveal key drivers of temporal dynamics in a synthetic yeast community

Multiple-model machine learning identifies potential functional genes in dilated cardiomyopathy

Contact Info

Product

Resources

About