Biology subjects

Sperlea, T.

Publications and source records attributed to Sperlea, T..

8 recordsLinked to original sources

bioRxiv2026

Semi-Annual Cycles in the Biotic Communities of Temperate Aquatic Habitats

Recurring patterns in biosphere dynamics are anchored in daily and seasonal oscillations in abiotic variables driven by Earths obliquity, rotation, and orbit. While circadian and annual biotic cycles are well studied, persistent supra- or subannual cycles in biotic systems are rarely documented globally. Here, we apply a machine learning approach to DNA metabarcoding time series and detect a biotic semi-annual cycle expressed across aquatic communities in temperate regions across taxonomic domains. We propose that this dynamic reflects a semi-annual mode in insolation and is suppressed under conditions of limited nutrients or sunlight. Our results suggest photoautotrophs are central for the aetiology of the biotic SAM, while demonstrating that it is a community-level phenomena not attributable to single species. The regularity of the biotic SAM suggests value for anticipating less predictable ecological events, including phytoplankton blooms. Overall, our results highlight Earth system-scale forcing of local dynamics and reinforce coupling patterns.

ecology↗

bioRxiv2026

Enhancing the Understanding of Environmental Microbiomes through Topic Modeling: A Quantitative and Qualitative Analysis

BackgroundUnderstanding ecosystem dynamics is essential for assessing ecosystem health, yet remains challenging due to complex biotic and abiotic interactions. Microbial communities are valuable indicators of environmental change, but the high dimensionality of microbiome data requires advanced analytical methods. This study explores the use of topic modeling (TM), an unsupervised machine learning approach initially designed for text analysis, to analyze microbiome data from the dynamic Warnow Estuary on the southern Baltic Sea coast. ResultsWe applied TM to estuarine microbiome data and compared its performance to traditional dimensionality reduction methods, Principal Component Analysis (PCA) and Principal Coordinate Analysis (PCoA). Quantitative results indicate that TM performs comparably to conventional approaches in preserving ecological and functional information, and in certain aspects even superior. In addition, we show qualitatively that NNMF, a TM method, captures latent patterns in the data providing an interpretable perspective on the microbiome. In this exploratory framework, NNMF suggested five distinct sub-communities within the estuary that appear to follow a seasonal succession influenced by freshwater inflow. These sub-communities were associated with specific ranges of salinity and temperature and showed distinct taxonomic profiles, with shared characteristics across the estuarine system. ConclusionsOur findings suggest that TM is a useful tool for exploring complex environmental microbiome datasets, offering a complementary perspective that can provide additional ecological insights. TMs ability to highlight coherent microbial community patterns indicates its promise for supporting environmental monitoring and informing targeted ecosystem management in dynamic habitats, though further studies are needed to fully assess its applicability.

ecology↗

bioRxiv2025

Bacterial community adaptation after freshwater and seawater coalescence

Microbial community coalescence, the merging of entire microbial communities, is common across ecosystems, particularly in estuaries where freshwater and seawater mix. The complexity of these habitats makes the in situ study of community dynamics after coalescence challenging, highlighting the need for controlled experiments to unravel the factors influencing the coalescence in the estuary. To study these processes, we combined natural freshwater and seawater bacterial communities at five different mixing ratios and incubated them in parallel microcosms containing freshwater or seawater incubation media. Forty mixed communities were tracked over six passages using Nanopore full-length 16S rRNA gene sequencing. In the original field samples, freshwater hosted more diverse communities than seawater. While the communities were structurally distinct, shared bacterial families accounted for approximately 95% of total reads. Many low-abundance taxa were lost upon laboratory incubation, while potentially faster-growing ones were enriched. We found that the coalescence outcome was strongly shaped by the incubation media, whereas the mixing ratio had a minor influence. Mixed communities converged toward the source community native to the incubation media, with increasing similarity at a higher source community proportion. In freshwater, a 25% inoculum of the freshwater community was sufficient to re-establish a near-native freshwater community, whereas in seawater, similarity to the native seawater community depended on the inoculation ratio. Network analysis showed a tightly connected module of seawater families, reflecting their shared habitat preference or cooperation, whereas freshwater families were more loosely connected. We also observed that most families were unaffected by mixing ratios or temporal dynamics, with only a few showing mixing ratio dependence. For instance, the freshwater family Comamonadaceae and seawater families Marinomonadaceae and Pseudoalteromonadaceae were dominant in their respective native environments, and increased in mixed communities in proportion to their initial source proportion. Overall, environmental filtering had a stronger impact than the mixing ratio of source communities on coalesced communities, with habitat-specific taxa further modulating the outcome. These findings advanced our understanding of microbial responses to coalescence and provided insights into microbial community assembly in dynamic estuarine systems. HighlightsO_LIEnvironmental filtering outweighs the microbial source community ratio in shaping coalescence outcomes. C_LIO_LIAsymmetric resilience: freshwater communities require a lower inoculum to re-establish than seawater communities. C_LIO_LIModularity of the seawater source community was observed during coalescence. C_LIO_LIFull-length 16S rRNA gene profiling with a custom dual-barcoding Nanopore protocol enables cost-effective bacterial community tracking. C_LIO_LIControlled coalescence experiments offer mechanistic insights into estuarine microbial community transitions. C_LI

ecology↗

bioRxiv2025

Predicting Antifouling Paint Particle Contamination based on 16S rRNA Gene Sequencing Data using Random Forest-Based Machine Learning

Antifouling paints often contain biocides designed to inhibit biological growth, and antifouling paint particles (APPs) have been previously shown to affect microbial communities in sediment. Given typical methods for monitoring for APP presence can be specialised and challenging, alternative methods using simple, standardised and universal approaches, such as 16S rRNA amplicon sequencing, would be highly valuable. This study uses a field-based mesocosm approach to train a random forest-based (supervised) machine learning model to predict APPs presence and concentration in sediment based on 16S microbial community data. The model correctly predicted 100% APP-presence samples and 83.3% APP-absence samples in the incubation testset, although the model could not correctly predict APP concentration with sufficient accuracy. To determine real-world applicability of the model, samples from 14 different sites along the Baltic Sea coastline and Warnow estuary in NE Germany were collected and APP-presence was pre-determined using SEM-EDX spectroscopy. The model correctly assigned APP-absence status to all APP-absent sites, and correctly assigned 3 of 5 APP-contaminated sites as having APP presence. As such, these results confirm it is possible to predict APPs in sediment based on the microbial community, and serves as a proof-of-concept for the further development of machine learning-based predictive tools for environmental monitoring.

microbiology↗

bioRxiv2025

Phylogeny-aware comparative genomics of Vibrio vulnificus links genetic traits to pathogenicity

Vibrio vulnificus is a natural inhabitant of coastal brackish waters worldwide and an opportunistic pathogen that can cause severe infections and septicemia through seafood consumption or wound exposure. Due to global warming, its abundance is increasing at high latitudes. While the species harbors diverse virulence factors, its precise disease mechanisms remain unclear. Comparative genomics between clinical and environmental isolates can help identify key virulence genes, but the limited availability of genomes from environmental isolates has hindered progress. In this study, we sequenced the genomes of 82 V. vulnificus isolates from water, sediment, and seagrass along the Baltic Sea coast and complemented with published genomes from 208 clinical and 117 globally distributed isolates for comparative analysis. Phylogenetic reconstruction confirmed four major lineages, with Baltic Sea strains confined to lineage L2 and L4, while clinical and environmental strains were distributed across all lineages. This suggests that the phylogenetic structure of V. vulnificus reflects adaptation to environmental conditions rather than pathogenicity. Using the PhyloBOTL pipeline developed here, we identified 128 orthologs significantly enriched in clinical isolates, grouped into 36 co-localization clusters based on proximity in the genomes. These included genes linked to virulence, such as those for capsular polysaccharide synthesis and biofilm formation, as well as previously unrecognized candidates, including chaperone-usher pilus biosynthesis, spermidine synthesis, Type VI secretion effectors, and an RTX toxin-like protein. Several of the clinically enriched gene clusters have been independently lost in three V. vulnificus clades, suggesting convergent evolution and a distinct ecological niche shared by these claded. Finally, we used the clinically enriched genes to design PCR primers for detecting and monitoring pathogenic V. vulnificus strains, providing a valuable tool for surveillance and public health efforts.

microbiology↗

bioRxiv2021

The relationship between land cover and microbial community composition in European lakes

Microbes such as bacteria, archaea, and protists are essential for element cycling and ecosystem functioning, but many questions central to the understanding of the role of microbes in ecology are still open. Here, we analyze the relationship between lake microbiomes and the land cover surrounding the lakes. By applying machine learning methods, we quantify the covariance between land cover categories and the microbial community composition recorded in the largest amplicon sequencing dataset of European lakes available to date. We identify microbial bioindicators for these land cover categories. Combining land cover and physico-chemical bioindicators identified from the same amplicon sequencing dataset, we develop two novel similarity metrics that facilitate insights into the ecology of the lake microbiome. We show that the bioindicator network, i.e., the graph linking OTUs indicative of the same environmental parameters, corresponds to microbial co-occurrence patterns. Taken together, we demonstrate the strength of machine learning approaches to identify correlations between microbial diversity and environmental factors, potentially opening new approaches to integrate environmental molecular diversity into monitoring and water quality assessments.

ecology↗

bioRxiv2019

γBOriS: Identification of Origins of Replication in Gammaproteobacteria using Motif-based Machine Learning

The biology of bacterial cells is, in general, based on the information encoded on circular chromosomes. Regulation of chromosome replication is an essential process which mostly takes place at the origin of replication (oriC). Identification of high numbers of oriC is a prerequisite to enable systematic studies that could lead to insights of oriC functioning as well as novel drug targets for antibiotic development. Current methods for identyfing oriC sequences rely on chromosome-wide nucleotide disparities and are therefore limited to fully sequenced genomes, leaving a superabundance of genomic fragments unstudied. Here, we present{gamma} BOriS (Gammaproteobacterial oriC Searcher), which accurately identifies oriC sequences on gammaproteobacterial chromosomal fragments by employing motif-based DNA classification. Using{gamma} BOriS, we created BOriS DB, which currently contains 25,827 oriC sequences from 1,217 species, thus making it the largest available database for oriC sequences to date.

bioinformatics↗

bioRxiv2019

Deep Learning on Chaos Game Representation for Proteins

Classification of protein sequences is one big task in bioinformatics and has many applications. Different machine learning methods exist and are applied on these problems, such as support vector machines (SVM), random forests (RF), and neural networks (NN). All of these methods have in common that protein sequences have to be made machine-readable and comparable in the first step, for which different encodings exist. These encodings are typically based on physical or chemical properties of the sequence. However, due to the outstanding performance of deep neural networks (DNN) on image recognition, we used frequency matrix chaos game representation (FCGR) for encoding of protein sequences into images. In this study, we compare the performance of SVMs, RFs, and DNNs, trained on FCGR encoded protein sequences. While the original chaos game representation (CGR) has been used mainly for genome sequence encoding and classification, we modified it to work also for protein sequences, resulting in n-flakes representation, an image with several icosagons.\n\nWe could show that all applied machine learning techniques (RF, SVM, and DNN) show promising results compared to the state-of-the-art methods on our benchmark datasets, with DNNs outperforming the other methods and that FCGR is a promising new encoding method for protein sequences.

bioinformatics↗