RiboTaxa: combined approaches for rRNA genes taxonomic resolution down to the species level from metagenomics data revealing novelties

DeepMPTB: a vaginal microbiome-based deep neural network as artificial intelligence strategy for efficient preterm birth prediction

In recent decades, preterm birth (PTB) has become a significant research focus in the healthcare field, as it is a leading cause of neonatal mortality worldwide. Using five independent study cohorts including 1290 vaginal samples from 561 pregnant women who delivered at term (n = 1029) or prematurely (n = 261), we analysed vaginal metagenomics data for precise microbiome structure characterization. Then, a deep neural network (DNN) was trained to predict term birth (TB) and PTB with an accuracy of 84.10% and an area under the receiver operating characteristic curve (AUROC) of 0.875 ± 0.11. During a benchmarking process, we demonstrated that our DL model outperformed seven currently used machine learning algorithms. Finally, our results indicate that overall diversity of the vaginal microbiota should be taken in account to predict PTB and not specific species. This artificial-intelligence based strategy should be highly helpful for clinicians in predicting preterm birth risk, allowing personalized assistance to address various health issues. DeepMPTB is open source and free for academic use. It is licensed under a GNU Affero General Public License 3.0 and is available at https://deepmptb.streamlit.app/ . Source code is available at https://github.com/oschakoory/DeepMPTB and can be easily installed using Docker ( https://www.docker.com/ ).

Mock community experiments can inform on the reliability of eDNA metabarcoding data: a case study on marine phytoplankton

Environmental DNA metabarcoding is increasingly implemented in biodiversity monitoring, including phytoplankton studies. Using 21 mock communities composed of seven unicellular diatom and dinoflagellate algae, assembled with different composition and abundance by controlling the number of cells, we tested the accuracy of an eDNA metabarcoding protocol in reconstructing patterns of alpha and beta diversity. This approach allowed us to directly evaluate both qualitative and quantitative metabarcoding estimates. Our results showed non-negligible rates (17-25%) of false negatives (i.e., failure to detect a taxon in a community where it was included), for three taxa. This led to a statistically significant underestimation of metabarcoding-derived alpha diversity (Wilcoxon p = 0.02), with the detected species richness being lower than expected (based on cell numbers) in 8/21 mock communities. Considering beta diversity, the correlation between metabarcoding-derived and expected community dissimilarities was significant but not strong (R2 = 0.41), indicating suboptimal accuracy of metabarcoding results. Average biovolume and rDNA gene copy number were estimated for the seven taxa, highlighting a potential, though not exhaustive, role of the latter in explaining the recorded biases. Our findings highlight the importance of mock communities for assessing the reliability of phytoplankton eDNA metabarcoding studies and identifying their limitations.

Temporal Changes in the Skin Microbiome of Epidermolysis Bullosa Patients following the Application of Wound Dressings

OBJECTIVE

Epidermolysis bullosa (EB) is a group of rare hereditary skin disorders characterized by the formation of painful blisters, erosions, and ulcers. In addition, the wounds can easily become infected with different pathogens. Therefore, the dynamics in the microbial populations across the various stages of EB can shed light on pathophysiology, the effect of treatment, and the factors involved in its recovery, but they are understudied. We thus sought to characterize the skin microbiome among patients with EB over time.

METHODS

A prospective study conducted in the pediatric dermatology clinic at Soroka Medical Center, Beer-Sheva, Israel. Children (0-18) with simplex and recessive dystrophic EB were sampled at two different time points: before a therapeutic regimen and 90 days (±14 days) later. Samples were obtained from lesional skin (wound), healthy, non-lesional skin, and seborrheic skin (forehead). Samples were subject to 16S rRNA amplicon sequencing. Analyses performed included comparisons of relative abundance at the phyla and genera taxonomic levels, alpha and beta diversity comparisons, and differential abundance.

RESULTS

32 children with EB were enrolled, for whom 192 skin microbiome samples were obtained. Lesional skin samples harbored significantly less Bacteroidota and Fusobacteriota before the initiation of treatment. Following topical dressing, we observed more Firmicutes and less Proteobacteria in lesional skin samples than healthy and seborrheic skin samples. In addition, Staphylococcus was significantly more abundant in lesional samples than in non-lesional and seborrheic samples following treatment.

CONCLUSIONS

Our study recaptured the reduced bacterial diversity and increased staphylococcal carriage in EB patients, showing a potential effect of topical dressing either directly on the wound microbiome or indirectly through the contribution towards skin healing. The detection of Firmicutes in general, and S. aureus specifically, commensurate with the application of a wound dressing may warrant the use of additional treatment methods to facilitate wound healing. Future studies in these patients should prospectively correlate the temporal changes in the microbiome associated with various treatment modalities in order to optimize the care of EB patients.