SuSha: A multi-model ensemble learning framework for predicting microbial salinity adaptation

Abstract

Current research on microbial salinity adaptation faces substantial challenges, including the limited predictive accuracy of traditional single-gene models and difficulty in dissecting systemic biological responses to salinity stress in complex natural habitats. To overcome these bottlenecks, the multi-model ensemble learning tool SuSha, which leverages genome-wide amino acid composition features, was developed. By extracting features from the whole-genome data of 123 bacterial and archaeal species with well-defined salinity adaptations, a 24-dimensional feature vector was constructed, comprising the frequencies of 20 standard amino acids and four aggregated functional categories. Based on this, an ensemble model was developed by integrating algorithms such as random forest, bagging, and extra trees. Five-fold cross-validation demonstrated that this 24-dimensional feature-based ensemble model achieved a global accuracy of 0.765 and an area under the curve of 0.941, significantly outperforming individual baseline models. Furthermore, the model was externally validated using 2678 metagenomic samples from six global regions, encompassing freshwater, marine, and hypersaline habitats. SuSha exhibited high robustness, ecological consistency across diverse salinity gradients, and a classification accuracy of over 90% for extreme halophiles, particularly within the extreme halophilic range. By enabling high-precision genotype-to-phenotype predictions using a habitat-adaptive algorithm-switching strategy, SuSha provides a robust computational framework for inferring the physiological potential of uncultivated microorganisms and mining microbial resources in extreme environments.

Keywords

Multi-model ensemble learning / Amino acid composition features / Salinity adaptation prediction / Extreme environment microbiomes / Uncultivated microorganisms

References

[1] L. Shen, Y. Liu, L. Chen, T. Lei, P. Ren, M. Ji, W. Song, H. Lin, W. Su, S. Wang, M. Rooman, F. Pucci, Genomic basis of environmental adaptation in the widespread poly—extremophilic exiguobacterium group , ISME J. 18 (2024) wrad020, doi: 10.1093/ismejo/wrad020.

[2] G. Merlino, A. Barozzi, G. Michoud, D.K. Ngugi, D. Daffonchio, Microbial ecology of deep—sea hypersaline anoxic basins, FEMS Microbiol. Ecol. 94 (2018) fiy085, doi: 10.1093/femsec/fiy085.

[3] L.R. Thompson, J.G. Sanders, D. McDonald, A. Amir, J. Ladau, K.J. Locey, et al., A communal catalogue reveals Earth’s multiscale microbial diversity, Nature 551 (2017) 457-463, doi: 10.1038/nature24621.

[4] K.G. Lloyd, A.D. Steen, J. Ladau, J. Yin, L. Crosby, Phylogenetically novel uncultured microbial cells dominate earth microbiomes, mSystems 3 (2018) e00055—18, doi: 10.1128/mSystems.00055-18.

[5] A.D. Steen, A. Crits—Christoph, P. Carini, K.M. DeAngelis, N. Fierer, K.G. Lloyd, J.C. Thrash, High proportions of bacteria and archaea across most biomes remain uncultured, ISME J. 13 (2019) 3126-3130, doi: 10.1038/s41396-019-0484-y.

[6] M. Vaish, A. Price—Whelan, T. Reyes—Robles, J. Liu, A. Jereen, S. Christie, F. Alonzo, M.A. Benson, V.J. Torres, T.A. Krulwich, Roles of Staphylococcus aureus Mnh1 and Mnh2 antiporters in salt tolerance, alkali tolerance, and pathogenesis, J. Bacteriol. 200 (2018) e00611—17, doi: 10.1128/JB.00611-17.

[7] C.D. Vavourakis, A.—S. Andrei, M. Mehrshad, R. Ghai, D.Y. Sorokin, G. Muyzer, A metagenomics roadmap to the uncultured genome diversity in hypersaline soda lake sediments, Microbiome 6 (2018) 168, doi: 10.1186/s40168-018-0548-7.

[8] N. Gunde—Cimerman, A. Plemenitaš, A. Oren, Strategies of adaptation of microorganisms of the three domains of life to high salt concentrations, FEMS Microbiol. Rev. 42 (2018) 353-375, doi: 10.1093/femsre/fuy009.

[9] V. Sabarly, O. Bouvet, J. Glodt, O. Clermont, D. Skurnik, L. Diancourt, D. De Vienne, E. Denamur, C. Dillmann, The decoupling between genetic structure and metabolic phenotypes in Escherichia coli leads to continuous phenotypic diversity , J. Evol. Biol. 24 (2011) 1559-1571, doi: 10.1111/j.1420-9101.2011.02287.x.

[10] S. Freilich, A. Kreimer, E. Borenstein, U. Gophna, R. Sharan, E. Ruppin, Decoupling environment—dependent and independent genetic robustness across bacterial species, PLoS Comput. Biol. 6 (2010) e1000690, doi: 10.1371/journal.pcbi.1000690.

[11] X.Z. Dong, Y.G. Zhou, H.H. Zhu, M. Cai, H.L. Cui, X. Dai, X.Y. Gao, C.Y. Jiang, A.H. Li, Z.L. Liang, H.C. Liu, Q. Liu, Y. Liu, Z.P. Liu, L. Song, Y.H. Xin, Manual of Systematic Identification of Common Bacteria and Archaea, 1st ed., Science Press, Beijing, 2023.

[12] R.R. de la Haba, C. Sánchez—Porro, M.C. Marquez, A. Ventosa, Taxonomy of Halophiles, 1st ed., Springer, Tokyo, 2011.

[13] F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, V. Dubourg, J. Vanderplas, A. Passos, D. Cournapeau, M. Brucher, M. Perrot, E. Duchesnay, Scikit—learn: machine learning in Python, J. Mach. Learn. Res. 12 (2011) 2825-2830, doi: 10.5555/1953048.2078195.

[14] O. Chapelle, V. Vapnik, Y. Bengio, Model selection for small sample regression, Mach. Learn 48 (2002) 9-23, doi: 10.1023/A:1013943418833.

[15] M. Fernández—Delgado, E. Cernadas, S. Barro, D. Amorim, Do we need hundreds of classifiers to solve real world classification problems? J. Mach. Learn. Res. 15 (2014) 3133-3181, doi: 10.5555/2627435.2697065.

[16] R. Leinonen, H. Sugawara, M. Shumway, on behalf of the International Nucleotide Sequence Database Collaboration, The sequence read archive, Nucleic. Acids. Res. 39 (2011) D19-D21, doi: 10.1093/nar/gkq1019.

[17] D. Li, C.—M. Liu, R. Luo, K. Sadakane, T.—W. Lam, MEGAHIT: an ultra—fast single—node solution for large and complex metagenomics assembly via succinct de Bruijn graph , Bioinformatics 31 (2015) 1674-1676, doi: 10.1093/bioinformatics/btv033.

[18] H. Li, Minimap2: pairwise alignment for nucleotide sequences, Bioinformatics 34 (2018) 3094-3100, doi: 10.1093/bioinformatics/bty191.

[19] S. Pan, X.—M. Zhao, L.P. Coelho, SemiBin2: self—supervised contrastive learning leads to better MAGs for short— and long—read sequencing, Bioinformatics 39 (2023) i21-i29, doi: 10.1093/bioinformatics/btad209.

[20] A. Chklovski, D.H. Parks, B.J. Woodcroft, G.W. Tyson, CheckM2: a rapid, scalable and accurate tool for assessing microbial genome quality using machine learning, Nat. Methods 20 (2023) 1203-1212, doi: 10.1038/s41592-023-01940-w.

[21] T. Seemann, Prokka: rapid prokaryotic genome annotation, Bioinformatics 30 (2014) 2068-2069, doi: 10.1093/bioinformatics/btu153.

[22] K.M. Sujon, R. Hassan, K. Choi, M.A. Samad, Accuracy, precision, recall, f1—score, or MCC? Empirical evidence from advanced statistics, ML, and XAI for evaluating business predictive models, J. Big Data 12 (2025) 268, doi: 10.1186/s40537-025-01313-4.

[23] R. Ghai, L. Pašić, A.B. Fernández, A.—B. Martin—Cuadrado, C.M. Mizuno, K.D. McMahon, R.T. Papke, R. Stepanauskas, B. Rodriguez—Brito, F. Rohwer, C. Sánchez—Porro, A. Ventosa, F. Rodríguez—Valera, New abundant microbial groups in aquatic hypersaline environments, Sci. Rep. 1 (2011) 135, doi: 10.1038/srep00135.

[24] R.L. Couto—Rodríguez, R. Montalvo—Rodríguez, Temporal analysis of the microbial community from the crystallizer ponds in Cabo Rojo, Puerto Rico, using metagenomics, Genes 10 (2019) 422, doi: 10.3390/genes10060422.

[25] U. Battumur, M. Lee, G.S. Bae, C.—H. Kim, Isolation and characterization of a new methanoculleus bourgensis strain KOR—2 from the rumen of Holstein steers , Asian—Australas J. Anim. Sci. 32 (2019) 241-248, doi: 10.5713/ajas.18.0409.