lgpr: An interpretable nonparametric method for inferring covariate effects from longitudinal Data

Localized coevolution between microbial predator and prey alters community-wide gene expression and ecosystem function

Closely interacting microbial species pairs (e.g., predator and prey) can become coadapted via reciprocal natural selection. A fundamental challenge in evolutionary ecology is to untangle how coevolution in small species groups affects and is affected by biotic interactions in diverse communities. We conducted an experiment with a synthetic 30-species bacterial community where we experimentally manipulated the coevolutionary history of a ciliate predator and one bacterial prey species from the community. Altering the coevolutionary history of the focal prey species had little effect on community structure or carrying capacity in the presence or absence of the coevolved predator. However, community metabolic potential (represented by per-cell ATP concentration) was significantly higher in the presence of both the coevolved focal predator and prey. This ecosystem-level response was mirrored by community-wide transcriptional shifts that resulted in the differential regulation of nutrient acquisition and surface colonization pathways across multiple bacterial species. Our findings show that the disruption of localized coevolution between species pairs can reverberate through community-wide transcriptional networks even while community composition remains largely unchanged. We propose that these altered expression patterns may signal forthcoming evolutionary and ecological change.

Deep longitudinal multi-omics analysis of Bordetella pertussis cultivated in bioreactors highlights medium starvations and transitory metabolisms, associated to vaccine antigen biosynthesis variations and global virulence regulation

Bordetella pertussis is the bacterial causative agent of whooping cough, a serious respiratory illness. An extensive knowledge on its virulence regulation and metabolism is a key factor to ensure pertussis vaccine manufacturing process robustness. The aim of this study was to refine our comprehension of B. pertussis physiology during in vitro cultures in bioreactors. A longitudinal multi-omics analysis was carried out over 26 h small-scale cultures of B. pertussis. Cultures were performed in batch mode and under culture conditions intending to mimic industrial processes. Putative cysteine and proline starvations were, respectively, observed at the beginning of the exponential phase (from 4 to 8 h) and during the exponential phase (18 h 45 min). As revealed by multi-omics analyses, the proline starvation induced major molecular changes, including a transient metabolism with internal stock consumption. In the meantime, growth and specific total PT, PRN, and Fim2 antigen productions were negatively affected. Interestingly, the master virulence-regulating two-component system of B. pertussis (BvgASR) was not evidenced as the sole virulence regulator in this in vitro growth condition. Indeed, novel intermediate regulators were identified as putatively involved in the expression of some virulence-activated genes (vags). Such longitudinal multi-omics analysis applied to B. pertussis culture process emerges as a powerful tool for characterization and incremental optimization of vaccine antigen production.

BaySyn: Bayesian Evidence Synthesis for Multi-system Multiomic Integration

The discovery of cancer drivers and drug targets are often limited to the biological systems - from cancer model systems to patients. While multiomic patient databases have sparse drug response data, cancer model systems databases, despite covering a broad range of pharmacogenomic platforms, provide lower lineage-specific sample sizes, resulting in reduced statistical power to detect both functional driver genes and their associations with drug sensitivity profiles. Hence, integrating evidence across model systems, taking into account the pros and cons of each system, in addition to multiomic integration, can more efficiently deconvolve cellular mechanisms of cancer as well as learn therapeutic associations. To this end, we propose BaySyn - a hierarchical Bayesian evidence synthesis framework for multi-system multiomic integration. BaySyn detects functionally relevant driver genes based on their associations with upstream regulators using additive Gaussian process models and uses this evidence to calibrate Bayesian variable selection models in the (drug) outcome layer. We apply BaySyn to multiomic cancer cell line and patient datasets from the Cancer Cell Line Encyclopedia and The Cancer Genome Atlas, respectively, across pan-gynecological cancers. Our mechanistic models implicate several relevant functional genes across cancers such as PTPN6 and ERBB2 in the KEGG adherens junction gene set. Furthermore, our outcome model is able to make higher number of discoveries in drug response models than its uncalibrated counterparts under the same thresholds of Type I error control, including detection of known lineage-specific biomarker associations such as BCL11A in breast and FGFRL1 in ovarian cancers. All our results and implementation codes are freely available via an interactive R Shiny dashboard at tinyurl.com/BaySynApp. The supplementary materials are available online at tinyurl.com/BaySynSup.