Shekhar, A. | Search bioRxiv

bioRxiv2026

Deep soil water as a dominant source for spruce water uptake in a subalpine forest: evidence from multi-year isotope data

- Norway spruce (Picea abies) dominates many European mountain forests, yet their seasonal water uptake strategies in high-elevation mono-specific natural stands remain poorly understood. We quantified contributions of shallow (0-10 cm) and deep (50-70 cm) soil layers to tree water uptake over three consecutive growing seasons (2020-2022) using stable water isotopes and Bayesian mixing analysis. - Contrary to the prevailing view of spruce as a shallow-rooted species relying primarily on water from the upper 10-20 cm of soil, our results showed more than 50% water uptake from deeper soil (50-70 cm), with deeper soil contributions crossing 80% in 2020. - During the dry and warm summer of 2022, positive soil recharge and elevated atmospheric demand increased evapotranspiration, with spruce trees taking up recently infiltrated rainfall from different soil depths, including >50% uptake from deeper layers. - Spruce water uptake shifted from cold-season-recharged soil water early in the growing season to warm-season precipitation in late summer. The timing of this shift in mid-summer can be explained by soil water recharge from recent rainfall infiltrated into the entire soil profile. This reliance on summer precipitation increases vulnerability of mono-specific spruce stands to more frequent droughts and heat waves under future climate change.

plant biology↗

bioRxiv2026

Systematic functional dissection of germline noncoding risk variants impacting clonal hematopoiesis

Clonal hematopoiesis of indeterminate potential (CHIP) is a precursor condition characterized by the expansion of blood cell clones harboring somatic mutations originating in hematopoietic stem cells (HSCs). Since individuals with CHIP face a high risk of developing myeloid malignancies, targeting CHIP clones could provide a viable strategy for leukemia prevention. Despite its clinical significance, the mechanisms underlying CHIP predisposition and progression remain poorly understood. Recent genome wide association studies (GWAS) have identified several non-coding genetic loci that are strongly associated with CHIP; however, their underlying mechanisms still remain unknown. We hypothesize that risk variants in these non-coding loci modulate enhancer elements active in HSCs. To test this, we selected 1,374 non-coding variants from 51 loci associated for CHIP risk in the UK Biobank and screened them for regulatory activity using a Massively Parallel Reporter Assay (MPRA). We performed our lentiviral MPRA screen in MUTZ-3 cells, a human hematopoietic cell line relevant to HSCs, which express CD34 surface marker and are dependent on HSC-specific transcription factors. Using a MPRA library of [~]73,000 constructs in CD34+ fraction of MUTZ-3 cells, we identified 87 variants representing 32 GWAS loci. We used targeted genome editing to demonstrate endogenous enhancer activity across 3 MPRA variants that affect the transcription of NKD2, FLT3, and MSI2. Our functional studies on MSI2 indicate that presence of higher levels of MSI2 mediated by CHIP risk allele enhances the clonal expansion of TET2 knockout hematopoietic stem and progenitor cells, providing a mechanistic link whereby non-coding genetic variants can influence the expansion of mutant CHIP clones.

genetics↗

bioRxiv2025

A Computational Pipeline for Glioblastoma Vaccine Development: Integrating Novel Omics-Driven OIP5 Target Discovery to Create a Deep Learning-Based Immunogenicity Framework for Personalized Immunotherapy

This work introduces a modular, open-source computational pipeline for glioblastoma (GBM) vaccine design that integrates omics-based OIP5 target discovery with a deep learning framework for epitope immunogenicity prediction. Building upon conventional affinity-based predictors such as NetMHCpan, our Tumor Epitope Immunogenicity Pipeline (TEIP) incorporates biological, structural, and transcriptomic context to predict tumor-specific T-cell responses. Using curated immunogenic and non-immunogenic peptides from IEDB and GBM datasets, TEIP employs dual bidirectional LSTM encoders to represent peptide and HLA sequences, concatenated with auxiliary molecular features including proteasomal cleavage, TAP transport likelihood, gene expression, and mutation frequency. These features are fused into a probabilistic model that outputs peptide-specific immunogenicity scores and confidence estimates. Benchmark results across multiple HLA alleles demonstrate that TEIP outperforms NetMHCpan and MARIA, achieving an ROC-AUC of 0.89 and PR-AUC of 0.35. Feature importance analysis confirms that tumor-specific expression and peptide-MHC binding dominate predictive accuracy, validating the biological realism of the model. By combining mechanistic features with data-driven representation learning, TEIP enables fine-grained prioritization of neoantigens with translational potential. In proof-of-concept GBM studies, TEIP recovered known cancer-testis antigens such as OIP5, demonstrating its ability to identify clinically relevant vaccine targets. The pipeline is implemented using entirely open-access data and software to promote reproducibility and scalability. Collectively, TEIP provides a unified framework that connects multi-omics tumor characterization with deep learning-based antigen modeling, establishing a foundation for precision immunotherapy development in glioblastoma and beyond.

bioinformatics↗

bioRxiv2025

TEIP: A Compact, Open-Source Framework for Predicting Tumor Epitope Immunogenicity in Glioblastoma Cancer Using Deep Learning and Multi-Modal Biological Features

This work introduces a modular, open-source computational pipeline for glioblastoma (GBM) vaccine design that integrates omics-based OIP5 target discovery with a deep learning framework for epitope immunogenicity prediction. Building upon conventional affinity-based predictors such as NetMHCpan, our Tumor Epitope Immunogenicity Pipeline (TEIP) incorporates biological, structural, and transcriptomic context to predict tumor-specific T-cell responses. Using curated immunogenic and non-immunogenic peptides from IEDB and GBM datasets, TEIP employs dual bidirectional LSTM encoders to represent peptide and HLA sequences, concatenated with auxiliary molecular features including proteasomal cleavage, TAP transport likelihood, gene expression, and mutation frequency. These features are fused into a probabilistic model that outputs peptide-specific immunogenicity scores and confidence estimates. Benchmark results across multiple HLA alleles demonstrate that TEIP outperforms NetMHCpan and MARIA, achieving an ROC-AUC of 0.89 and PR-AUC of 0.35. Feature importance analysis confirms that tumor-specific expression and peptide-MHC binding dominate predictive accuracy, validating the biological realism of the model. By combining mechanistic features with data-driven representation learning, TEIP enables fine-grained prioritization of neoantigens with translational potential. In proof-of-concept GBM studies, TEIP recovered known cancer-testis antigens such as OIP5, demonstrating its ability to identify clinically relevant vaccine targets. The pipeline is implemented using entirely open-access data and software to promote reproducibility and scalability. Collectively, TEIP provides a unified framework that connects multi-omics tumor characterization with deep learning-based antigen modeling, establishing a foundation for precision immunotherapy development in glioblastoma and beyond.

cancer biology↗

bioRxiv2025

Obligate cross-feeding of metabolites is common in soil microbial communities

Many microorganisms are refractory to laboratory cultivation. One possible explanation, known as the great plate count anomaly, is metabolic dependencies among community members. However, systematic studies of these interactions in communities like soil are missing, hindering advances in understanding these ecologically important ecosystems. Here, we address this issue by systematically analysing 6,931 bacterial isolates of 27 soil microbial communities. We find that the growth of up to 50% of all community members depended essentially on supplementation with amino acids, vitamins, or nucleotides. In 73% of cases, supplementation with multiple amino acids was necessary. Genomic analysis of 62 strains revealed that accumulation of insertion sequences and specific gene loss was associated with the observed auxotrophies. Finally, genome-scale metabolic models, computational analyses, and cocultivation experiments demonstrated that other co-occurring genotypes complemented the metabolic needs of auxotrophs, thus facilitating their growth. Our results demonstrate that soil bacteria exist within integrated metabolic networks, which hampers their cultivation.

microbiology↗

bioRxiv2020

Wide-Field Calcium Imaging of Dynamic Cortical Networks During Locomotion

Motor behavior results in widespread activation of the cerebral cortex. Therefore, fully understanding the cerebral cortexs role in motor behavior requires a mesoscopic level description of the engaged cortical regions and their functional interactions. Mesoscopic imaging of Ca2+ fluorescence through transparent polymer skulls implanted on transgenic Thy1-GCaMP6f mice reveals widespread activation of the cerebral cortex during locomotion, including not only in primary motor and somatosensory regions but also in secondary motor, retrosplenial, and visual cortices. Using the time series of Ca2+ fluorescence from 28 regions (nodes) we found to be consistent across mice, we examined the changes in functional connectivity from rest to locomotion. At the initiation of locomotion, the most anterior nodes of the secondary motor cortex increase in correlation with other nodes, while other nodes decrease in correlation. Eigenvector centrality confirms these changes in functional connectivity. Directed Granger causality analysis reveals an increase in causal influence of anterior regions of secondary motor cortex on other dorsal cortical regions at the onset of locomotion. These results highlight the global changes in correlation, centrality, and causality occurring throughout the cerebral cortex between rest and locomotion and suggest that the premotor areas play an important role in organizing these changes.

neuroscience↗