Biology subjects

Patil, P.

Publications and source records attributed to Patil, P..

At least 19 recordsLinked to original sources

bioRxiv2026

ADMETron: An AI-driven SaaS platform for comprehensive ADMET prediction and compound prioritisation

ONTOSIGHT(R) ADMETron is a high-performance SaaS based AI platform designed for the rapid profiling and visualization of ADMET (Absorption, Distribution, Metabolism, Excretion, and Toxicity) properties. The platform integrates a highly interactive web interface with a robust predictive engine, enabling the batch processing of compounds for high-throughput virtual screening. The core engine employs an ensemble model that combines recurrent neural network (RNN)-derived embeddings from SMILES strings with physicochemical descriptors, which are fed into gradient boosting machines (GBMs). This architecture provides accurate predictions across 34 distinct ADMET endpoints, encompassing critical categories such as physicochemical properties, absorption, CYP450 inhibition, hERG toxicity, and mutagenicity. The platforms superior performance is quantitatively validated by its top-tier ranking on the Therapeutics Data Commons (TDC) ADMET Benchmark Group, demonstrating robustness and generalizability with notable results including 2nd place for Ames mutagenicity (AUROC 0.870) and 2nd place for LD50 (MAE 0.573). In addition to its predictive capabilities, ADMETron introduces a novel SAR analysis framework that enables real-time comparison of multiple compounds and approved drugs through an interactive radar graph visualization. Comparative evaluation against widely used online ADMET platforms demonstrated broader endpoint coverage, including pharmacokinetic, physicochemical, and medicinal chemistry assessments within a unified environment. The combination of benchmark-validated predictive performance, comprehensive ADMET profiling, and advanced visualization tools positions ADMETron as a next-generation platform for virtual screening, lead optimization, and data-driven decision-making in modern drug discovery (https://admetron.partex.ai/).

bioinformatics↗

bioRxiv2026

Proviral dynamics and HIV-1C viral diversity in the context of HIV-TB co-infection

BackgroundART effectively suppresses HIV replication and restores CD4+ T cells; however, long-lived HIV latent reservoirs enable viral persistence. Tuberculosis (TB) co-infection further impacts HIV latency and enhances viral replication. Given the high prevalence of latent TB infection (LTBI) in TB-endemic settings, understanding its impact on HIV biology is critical. Our study aims to investigate the influence of TB co-infection on HIV reservoir dynamics, viral diversity, and drug resistance mutations in ART-naive individuals. MethodologySamples from 90 ART-naive HIV-1C individuals, stratified based on IGRA and TB diagnosis, were used in this study. Plasma and PBMCs were isolated for viral RNA and DNA extraction respectively. Total proviral DNA was quantified using gag PCR. Full-length env and pol genes were amplified, purified and sequenced using ONT and Illumina platforms. Pol sequences were subjected to Drug Resistance Mutation (DRM) analysis via Stanford HIVdb with a minimum threshold mutation frequency of [≥]10%. Full length env sequences were used for phylogenetic analysis by aligning with Indian Subtype C reference sequence and phylogenetic tree was generated using ggplot2. ResultProviral load analysis showed no significant differences across HIV+LTBI-, HIV+LTBI+, and HIV+TB+ groups, although a trend toward higher levels was observed in HIV+TB+ individuals. Correlation analysis revealed distinct immune associations, with HIV+LTBI+ individuals showing positive correlations with activation and PD-1 expression. Longitudinal analysis of proviral loads demonstrated a modest decline in proviral load post-ART but remained persistent for up to 18-20 months following initiation of ART accompanied by low level ongoing viral replication. DRM analysis revealed a 33% prevalence in ART-naive individuals, with higher occurrence in HIV+LTBI+ group. Of the identified DRMs, 38% (5/13) and 71% (5/7) in sequences obtained from PBMC and plasma respectively were attributed to polymorphic mutations associated with Integrase strand transfer inhibitors (INSTIs). DRMs within plasma and PBMC derived viruses showed high concordance. Phylogenetic analysis of env sequences indicated overlapping viral populations between the 3 groups, with greater diversity in PBMCs compared to plasma. ConclusionThe study highlights that HIV reservoir dynamics, drug resistance, and viral diversity are significantly influenced by TB co-infection. While proviral loads were comparable, LTBI-associated immune activation and granuloma niches may have driven viral diversification and DRM emergence. High concordance between compartments and presence of transmitted resistance underscore the need for baseline screening, multi-compartment analysis, and sustained surveillance.

microbiology↗

bioRxiv2026

Mena (ENAH) Promotes KRAS-Driven Tumor Growth and Metastatic Progression in Pancreatic Ductal Adenocarcinoma

Metastatic pancreatic ductal adenocarcinoma (PDAC) remains incurable and is projected to become the second leading cause of cancer-related death by 2030. Despite therapeutic advances, median survival remains under one year. Activating KRAS mutations drive the majority of PDAC and underlie their highly aggressive behavior. Although emerging KRAS inhibitors show promise, resistance limits their clinical efficacy, underscoring the need to identify additional regulators of mutant KRAS signaling. We investigated the role of the actin-regulatory protein Mena (ENAH) in KRASG12D-driven PDAC using novel in vitro and orthotopic in vivo models with Mena overexpression and knockdown. Mena overexpression markedly increased primary tumor growth and spontaneous liver metastasis, whereas Mena knockdown delayed tumor onset and reduced metastatic burden. Mechanistically, Mena depletion significantly decreased AKT and ERK activity downstream of KRAS in a SHIP2-mediated manner, indicating that Mena enhances oncogenic signaling required for PDAC progression. These findings reveal Mena as a critical modulator of KRASG12D-driven pancreatic tumor growth and metastasis. By promoting key survival and proliferation pathways, Mena contributes to the aggressive phenotype characteristic of KRAS-mutant PDAC and represents a promising therapeutic target for patients with both locoregional and metastatic disease.

cancer biology↗

bioRxiv2025

TMEM Doorway Mediated Metastasis in Pancreatic Ductal Adenocarcinoma by Tie2 Signaling

Pancreatic ductal adenocarcinoma (PDAC) is almost invariably fatal due to early hematogenous dissemination that occurs before the primary tumor is clinically detectable, yet the cellular mechanism of tumor cell intravasation has remained unknown. Using multiphoton intravital imaging in autochthonous and orthotopic PDAC models, we demonstrate that intravasation occurs at Tumor Microenvironment of Metastasis (TMEM) doorways--tri-cellular structures comprising a MENA-expressing tumor cell, a Tie2 macrophage, and an endothelial cell in direct contact. These structures are abundant in human PDAC, enriched for Tie2 macrophages, and markedly reduced after neoadjuvant chemotherapy. Selective pharmacologic inhibition of Tie2 with rebastinib decreases TMEM-associated transient vascular openings, suppresses circulating and hepatic disseminated tumor cells, and--when combined with perioperative FOLFIRINOX after curative-intent resection--improves median survival in murine PDAC. These findings establish TMEM doorways as a common, druggable mechanism of intravasation across epithelial cancers and identify Tie2 macrophages as a therapeutic target to prevent metastatic seeding in PDAC, a disease with no anti-metastatic therapies. TMEM doorway-mediated intravasation in PDAC supports its role as a common gateway for hematogenous metastasis in carcinoma.

cancer biology↗

bioRxiv2025

Cyclin-dependent kinase 9 inhibitors as oncogene signaling modulators in combination with targeted therapy for the treatment of colorectal cancer

BackgroundColorectal cancer (CRC) is the second deadliest cancer worldwide and new treatment options are urgently needed. Cyclin dependent kinase 9 (CDK9) promotes aberrant RNA transcription in cancer and is a promising target for cancer therapies. MethodsUsing CRC cell lines as well as newly established patient-derived organoid models of CRC, we studied the clinically promising CDK9 inhibitors (AZD4573, BAY1125152/VIP152/enitociclib, and NVP2) to determine their therapeutic potential. We investigated the efficacy and mechanisms of action through cell growth and cytotoxicity assays, RNAseq, immunoblotting, and IHC. ResultsWe found CDK9 inhibitors to be highly potent against CRC, through suppression of proliferation and induction of apoptosis. Our results demonstrated CDK9 inhibitors to be broadly active against a set of CRC models derived from a diverse patient population. Our mechanistic studies showed significant suppression of the Mitogen-active protein kinase (MAPK) signaling pathway due to CDK9 inhibitor treatment, suggesting that CDK9 inhibitor efficacy could be enhanced when combined with MAPK pathway inhibitors. As proof-of-concept, we found that CDK9 inhibitors and MEK inhibitors could be combined to synergistically suppress CRC growth and survival. ConclusionsCDK9 inhibitors show promising activity against patient-derived models of CRC. MAPK signaling is particularly suppressed by CDK9 inhibitors. Combining CDK9 inhibitors and targeted therapy against MAPK signaling pathway may be a viable strategy worthy of further investigation preclinically and clinically. State of translational relevanceNovel therapies for chemotherapy-refractory CRC remain urgently needed. Rapid validation of efficacy in appropriate translational models can provide the necessary rationale to test new therapies in clinical trials. Using newly established patient-derived organoids, we validate the preclinical efficacy of transcriptional CDK9 inhibitors against CRC. This is a therapeutic class that has been scarcely evaluated for efficacy against CRC. We found that CDK9 inhibitors potently inhibit patient-derived organoid models of CRC. Our mechanistic work also reveals that clinically achievable concentrations of CDK9 inhibitor treatment results in suppression of MAPK signaling. As proof-of-concept of the potential actionability of our findings, we demonstrate that CDK9 inhibitors can be combined with MEK inhibitors to synergistically suppress CRC growth. Our study provides new patient-derived CRC models as a resource available to others for therapeutic testing, as well as a roadmap for novel hypotheses about synergistic combination treatments using CDK9 inhibitors.

cancer biology↗

bioRxiv2025

A rapid non-traditional approach for developing biologicals against drug- resistant bacteria and yeasts.

Development of effective, long-lasting antibiotics is a challenge due to the rate at which the pathogens acquire resistance, prompting the major pharma players to omit them from their portfolio. We present a non-traditional approach for rapidly developing anti-infective biologicals that is resilient to resistance. Antibody-drug conjugates were developed against three different pathogens Pseudomonas aeruginosa (ABG 14), Staphylococcus aureus (ABG 16) and Candida. albicans (ABG 07) that were extremely specific and effective in neutralising them irrespective of their drug resistance profiles. Neutralising camelid antibody fragment (VHH) were isolated from immunised camel libraries by phage display that effectively neutralised the pathogens at a MIC 99 of 125 {micro}g mL-1 in vitro. Antimicrobial peptides (AMP) were then conjugated to them by pathogen-specific cleavable linkers and the resultant molecules were 10 -20 times more effective due to a dual mode of action-the inhibitory action of the VHH on surface transporters and enzymes and the activity of the AMPs released by the pathogen surface proteases. Called AbTids, these molecules were extremely specific, stable in plasma and were activated in the presence of 104 to 105 CFU mL-1 of the pathogens and had an efficacy in the sub-micromolar range (6.25-12.5 {micro}g mL-1, 250 nM) inhibiting their growth within 2 hrs. of administration. They were economically and efficiently produced in microbial production system as a single chain fusion protein. One of the molecules ABG 14 was characterised further and found to be pathogen specific with negligible frequency of resistance, was non-toxic, had the capability to destroy biofilms, and cleared a systemic infection in a mouse with a carbapenem resistant strain of P. aeruginosa at a dose of 5 mg kg-1. This strategy can be used to generate new hits against medically important pathogens in a matter of weeks by simply shuffling the components, using different VHHs and AMPs, bypassing years of expensive drug development efforts that can potentially rejuvenate drug discovery efforts against emerging superbugs.

microbiology↗

bioRxiv2025

Exploring Synergies in Brain-Machine Interfaces: Compression vs. Performance

Individuals with severe neurological injuries often rely on assistive technologies, but current methods have limitations in accurately decoding multi-degree-of-freedom (DoF) movements. Intracortical brain-machine interfaces (iBMIs) use neural signals to provide a more natural control method, but currently struggle with higher-DoF movements--something the brain handles effortlessly. It has been theorized that the brain simplifies high-DoF movement through muscle synergies, which link multiple muscles to function as a single unit. These synergies have been studied using dimensionality reduction techniques like principal component analysis (PCA), non-negative matrix factorization (NMF), and demixed PCA (dPCA) and successfully used to reduce noise and improve offline decoder stability in non-invasive applications. However, their effectiveness in improving decoding and generalizability for implanted recordings across varied tasks is unclear. Here, we evaluated if brain and muscle synergies can enhance iBMI performance in non-human primates performing a two-DoF finger task. Specifically, we tested if PCA, dPCA, and NMF could compress and denoise brain and muscle data and improve decoder generalization across tasks. Our results showed that while all methods effectively compressed data with minimal loss in decoding accuracy, none improved performance through denoising. Additionally, none of the methods enhanced generalization across tasks. These findings suggest that while dimensionality reduction can aid data compression, alone it may not reveal the "true" control space needed to improve decoder performance or generalizability. Further research is required to determine whether synergies are the optimal control framework or if alternative approaches are required to enhance decoder robustness in iBMI applications. Significance StatementMany researchers believe that brain and muscle synergies represent a fundamental control strategy and could enhance brain-machine interface (BMI) decoding performance. These synergies, extracted through dimensionality reduction techniques, are thought to simplify complex neural data, improving the efficiency and accuracy of BMI systems. In our study, we evaluated brain and muscle synergies in a dexterous finger task. We found that while these synergies effectively compressed high-dimensional data, they did not improve performance through denoising or generalize well across different contexts. Instead, the highest performance was achieved when using all available data, suggesting that synergies, although useful for data compression, may not provide the "true" control space needed to enhance decoder robustness or adaptability in implanted BMI systems.

neuroscience↗

bioRxiv2024

Exploring the trade-off between deep-learning and explainable models for brain-machine interfaces

People with brain or spinal cord-related paralysis often need to rely on others for basic tasks, limiting their independence. A potential solution is brain-machine interfaces (BMIs), which could allow them to voluntarily control external devices (e.g., robotic arm) by decoding brain activity to movement commands. In the past decade, deep-learning decoders have achieved state-of-the-art results in most BMI applications, ranging from speech production to finger control. However, the black-box nature of deep-learning decoders could lead to unexpected behaviors, resulting in major safety concerns in real-world physical control scenarios. In these applications, explainable but lower-performing decoders, such as the Kalman filter (KF), remain the norm. In this study, we designed a BMI decoder based on KalmanNet, an extension of the KF that augments its operation with recurrent neural networks to compute the Kalman gain. This results in a varying "trust" that shifts between inputs and dynamics. We used this algorithm to predict finger movements from the brain activity of two monkeys. We compared KalmanNet results offline (pre-recorded data, n = 13 days) and online (real-time predictions, n = 5 days) with a simple KF and two recent deep-learning algorithms with state-of-the-art results: tcFNN and LSTM. KalmanNet achieved comparable or better results than other deep learning models in offline and online modes, relying on the dynamical model for stopping while depending more on neural inputs for initiating movements. We further validated this mechanism by implementing a heteroscedastic KF that used the same strategy, and it also approached state-of-the-art performance while remaining in the explainable domain of standard KFs. However, we also see two downsides to KalmanNet. KalmanNet shares the limited generalization ability of existing deep-learning decoders, and its usage of the KF as an inductive bias limits its performance in the presence of unseen noise distributions. Despite this trade-off, our analysis successfully integrates traditional controls and modern deep-learning approaches to motivate high-performing yet still explainable BMI designs.

neuroscience↗

bioRxiv2024

Statins attenuate Wnt/β-catenin signaling by targeting SATB family proteins in colorectal cancer

Colorectal cancer is the second leading cause of cancer-related deaths worldwide, highlighting the need for improved treatments and advanced molecular research. A recent therapeutic approach focuses on repurposing drugs to target dysregulated pathways involved in tumorigenesis. Among these, statins, commonly known for lowering cholesterol, have attracted attention for their potential anti-cancer properties. Here, we provide direct evidence for the same by assessing the impact of statin treatment on lipid, transcript, and protein levels. Our findings reveal that statins specifically target key components of the Wnt/{beta}-catenin pathway, a major factor in adenoma formation, including the SATB (Special AT-rich Binding protein) family proteins. While SATB1 is recognized as a regulator of tumorigenesis, particularly under Wnt signaling, SATB2 appears to exert an opposing role. We demonstrate that statin treatment reciprocally alters the expression pattern of these proteins. Furthermore, a human clinical trial evaluating statins as an anti-cancer therapy supports the hypothesis that differential expression of SATB proteins is crucial in tumorigenic outcomes. In conclusion, this modulation by statin treatment suggests promising new therapeutic avenues through drug repurposing.

cancer biology↗

bioRxiv2024

Induction of territorial behavior and dominance hierarchies in laboratory mice

Territorial behaviors comprise a set of coordinated actions and response patterns found across animal species that promote the exclusive access to resources. House mice are highly territorial with a subset of males consistently attacking and chasing competing males to expel them from their territories and performing urine marking behaviors to signal the extent of their territories. Natural variation in territorial behaviors within a mouse colony leads to the formation of dominance hierarchies in which subordinate males can reside within the territory of a dominant male. While the full repertoire of such territorial behaviors and hierarchies has been extensively studied in wild-derived mice in semi-natural enclosures, so far they have not been established in the smaller enclosures and with the genetically-defined laboratory strains required for the application of neural recording and manipulation methods. Here, we present a protocol to induce an extensive repertoire of territorial behaviors in small enclosures in laboratory mice, including a method for the simultaneous tracking of urine marking behavior in mouse pairs. Using this protocol we describe the emergence of robust dominant-subordinate hierarchies between pairs of CD1 outbred or CD1xB6 F1 hybrid mice, but unexpectedly not in C57BL/6 inbred animals. Our behavioral paradigm opens the door for neurocircuit studies of territorial behaviors and social hierarchy in the laboratory.

animal behavior and cognition↗

bioRxiv2023

Analysis of the Cross-Study Replicability of Tuberculosis Gene Signatures Using 49 Curated Transcriptomic Datasets

BackgroundTuberculosis (TB) is the leading cause of infectious disease mortality worldwide. Numerous blood-based gene expression signatures have been proposed in the literature as alternative tools for diagnosing TB infection. Ongoing efforts are actively focused on developing additional signatures in other TB-related contexts. However, the generalizability of these signatures to different patient contexts is not well-characterized. There is a pressing need for a well-curated database of TB gene expression studies for the systematic assessment of existing and newly developed TB gene signatures. ResultsWe built the curatedTBData, a manually-curated database of 49 TB transcriptomic studies. This data resource is freely available through GitHub and as an R Bioconductor package that allows users to validate new and existing biomarkers without the challenges of harmonizing heterogeneous studies. We also demonstrate the use of this data resource with cross-study comparisons for 72 TB gene signatures. For the comparison of subjects with active TB from healthy controls, 19 gene signatures had weighted mean AUC of 0.90 or greater, with the highest result of 0.94. In active TB disease versus latent TB infection, 7 gene signatures had weighted mean AUC of 0.90 or greater, with a maximum of 0.93. We also explore ensembling methods for averaging predictions from multiple gene signatures to significantly improve diagnostic ability beyond any single signature. ConclusionsThe curatedTBData data package offers a comprehensive resource of curated gene expression and clinically annotated data. It could be used to identify robust new TB gene signatures, to perform comparative analysis of existing TB gene signatures, and to develop alternative gene set scoring or ensembling methods, among other things. This resource will also facilitate the development of new signatures that are generalizable across cohorts or more applicable to specific subsets of patients (e.g. with rare comorbid conditions, etc.). We demonstrated that these blood-based gene signatures could distinguish patients with distinct TB outcomes; moreover, the combination of multiple gene signatures could improve the overall predictive accuracy in differentiating these subtypes, which point out an important aspect for the translation of genomics to clinical implementation.

genomics↗

bioRxiv2023

Gut feeling: Extent of virulence and antibiotic resistance genes in Helicobacter pylori and campylobacteria

BackgroundHelicobacter pylori, a member of campylobacteria, is the leading cause of chronic gastritis and gastric cancer. Virulence and antibiotic resistance of H. pylori are of great concern to public health. However, the relationship between virulence and antibiotic resistance genes in H. pylori in relation to other campylobacteria remains unclear. Materials and MethodsBy using the virulence and comprehensive antibiotic resistance databases, we explored all available 354 complete genomes of H. pylori and compared it with 90 species of campylobacteria for virulence and antibiotic resistance genes/proteins. ResultsOn average, H. pylori had 129 virulence genes, highest among Helicobacter spp. and 71 antibiotic resistance genes, one of the lowest among campylobacteria. Just 2.6% of virulence genes were shared by all campylobacterial members, whereas 9.4% were unique to H. pylori. The cytotoxin-associated genes (cags) seemed to be exclusive to H. pylori. Majority of the isolates from Asia and South America were cag2-negative and many antibiotic resistance genes showed isolate-specific patterns of occurrence. Just 15 (8.8%) antibiotic resistance genes, but 103 (66%) virulence genes including 25 cags were proteomically identified in H. pylori. Arcobacterial members showed large variation in the number of antibiotic resistance genes and there was a positive relation with the genome size. ConclusionLarge repository of antibiotic resistance genes in campylobacteria and a unique set of virulence genes might have important implications in shaping the course of virulence and antibiotic resistance in H. pylori.

bioinformatics↗

bioRxiv2023

Reversal of carbapenem resistance in Pseudomonas aeruginosa by camelid single domain antibody fragment (VHH) against the C4 dicarboxylate transporter

The COVID-19 pandemic has hastened the problem of nosocomial drug-resistant pathogens and is exerting a huge toll on hospitalized patients in critical care settings. Most small molecule antibiotics are susceptible to bacterial resistance mechanisms and constantly becoming ineffective leading to rapid shrinkage of the antibiotic armamentarium. Pseudomonas aeruginosa, one of the most common pathogens in hospital-born infection and considered critically dangerous by WHO and CDC, is extremely difficult to treat with frontline drugs, the carbapenems. In this study, we developed a camelid antibody fragment (VHH) library against whole P. aeruginosa and isolated a highly potent neutralizing hit (PsC23) that selectively targets P. aeruginosa . At 25 {micro}g/mL, PsC23 inhibited growth of the ATCC 27853 and a locally isolated carbapenem resistant MCC 50428 P. aeruginosa. The target of PsC23 is the C4 dicarboxylate transporter that transports C4 metabolites to the glyoxylate shunt during oxidative stress that is present in pathogens but not the human host. This ultimately results in the blockade of the shunt affecting bacterial energy transduction that leads to disruption of drug efflux. Interestingly, in a neutropenic mouse with MCC 50428 systemic infection, PsC23 in combination with meropenem completely reversed the drug resistance and eliminated the pathogens from the blood. PsC23 was stable in human serum and had no hemolytic or cytotoxic effect on human cells. Taken together, this VHH if co administered with the currently available carbapenems would reverse carbapenem resistance and could be used to effectively control P. aeruginosa in critical care settings.

molecular biology↗

bioRxiv2023

Structural Optimization of siRNA Conjugates for Albumin Binding Achieves Effective MCL1-Targeted Cancer Therapy

The high potential for therapeutic application of siRNAs to silence traditionally undruggable oncogenic drivers remains largely untapped due to the challenges of tumor cell delivery. Here, siRNAs were optimized for in situ binding to albumin through C18 lipid modifications to improve pharmacokinetics and tumor delivery. Systematic variation of siRNA conjugates revealed a lead structure with divalent C18 lipids each linked through three repeats of hexaethylene glycol connected by phosphorothioate bonds. Importantly, we discovered that locating the branch site of the divalent lipid structure proximally (adjacent to the RNA) rather than at a more distal site (after the linker segment) promotes association with albumin, while minimizing self-assembly and lipoprotein association. Comparison to higher albumin affinity (diacid) lipid variants and siRNA directly conjugated to albumin underscored the importance of conjugate hydrophobicity and reversibility of albumin binding for siRNA delivery and bioactivity in tumors. The lead conjugate increased tumor siRNA accumulation 12-fold in orthotopic mouse models of triple negative breast cancer over the parent siRNA. When applied for silencing of the anti-apoptotic oncogene MCL-1, this structure achieved approximately 80% MCL1 silencing in orthotopic breast tumors. Furthermore, application of the lead conjugate structure to target MCL1 yielded better survival outcomes in three independent, orthotopic, triple negative breast cancer models than an MCL1 small molecule inhibitor. These studies provide new structure-function insights on optimally leveraging siRNA-lipid conjugate structures that associate in situ with plasma albumin for molecular-targeted cancer therapy.

bioengineering↗

bioRxiv2023

Comparison of gene set scoring methods for reproducible evaluation of multiple tuberculosis gene signatures

RationaleMany blood-based transcriptional gene signatures for tuberculosis (TB) have been developed with potential use to diagnose disease, predict risk of progression from infection to disease, and monitor TB treatment outcomes. However, an unresolved issue is whether gene set enrichment analysis (GSEA) of the signature transcripts alone is sufficient for prediction and differentiation, or whether it is necessary to use the original statistical model created when the signature was derived. Intra-method comparison is complicated by the unavailability of original training data, missing details about the original trained model, and inadequate publicly-available software tools or source code implementing models. To facilitate these signatures replicability and appropriate utilization in TB research, comprehensive comparisons between gene set scoring methods with cross-data validation of original model implementations are needed. ObjectivesWe compared the performance of 19 TB gene signatures across 24 transcriptomic datasets using both re-rebuilt original models and gene set scoring methods to evaluate whether gene set scoring is a reasonable proxy to the performance of the original trained model. We have provided an open-access software implementation of the original models for all 19 signatures for future use. MethodsWe considered existing gene set scoring and machine learning methods, including ssGSEA, GSVA, PLAGE, Singscore, and Zscore, as alternative approaches to profile gene signature performance. The sample-size-weighted mean area under the curve (AUC) value was computed to measure each signatures performance across datasets. Correlation analysis and Wilcoxon paired tests were used to analyze the performance of enrichment methods with the original models. Measurement and Main ResultsFor many signatures, the predictions from gene set scoring methods were highly correlated and statistically equivalent to the results given by the original diagnostic models. PLAGE outperformed all other gene scoring methods. In some cases, PLAGE outperformed the original models when considering signatures weighted mean AUC values and the AUC results within individual studies. ConclusionGene set enrichment scoring of existing blood-based biomarker gene sets can distinguish patients with active TB disease from latent TB infection and other clinical conditions with equivalent or improved accuracy compared to the original methods and models. These data justify using gene set scoring methods of published TB gene signatures for predicting TB risk and treatment outcomes, especially when original models are difficult to apply or implement.

bioinformatics↗

bioRxiv2022

Prevalence and heterogeneity of antibiotic-resistant genes in Orientia tsutsugamushi and other rickettsial genomes

Despite a million infections every year and an estimated one billion people at risk, scrub typhus is regarded as a neglected tropical disease. The causative bacterium Orientia tsutsugamushi, a member of rickettsiae, seems to be intrinsically resistant to several classes of antibiotics. The emergence of antibiotic-resistant scrub typhus is likely to become a global public health concern. Yet, it is unknown as to how common antibiotic-resistant genes are in O. tsutsugamushi, and how variable these loci are among the genomes of rickettsiae. By using the comprehensive antibiotic resistance database, we explored 79 complete genomes from 24 species of rickettsiae for putative antibiotic-resistant loci. There were 244 unique antibiotic-resistant genes in rickettsiae. Both the total and unique antibiotic-resistant genes in O. tsutsugamushi were significantly less compared to other members of rickettsiae. However, antibiotic-resistant genes in O. tsutsugamushi genomes were more unique and highly variable. Many genes such as resistant versions of evgS, and vanS A/G were present in numerous copies. These results will have important implications in the context of antibiotic-resistant scrub typhus.

bioinformatics↗

bioRxiv2022

A Pairwise Imputation Strategy for Retaining Predictive Features When Combining Multiple Datasets

In the training of predictive models using high-dimensional genomic data, multiple studies worth of data are often combined to increase sample size and improve generalizability. A drawback of this approach is that there may be different sets of features measured in each study due to variations in expression measurement platform or technology. It is often common practice to work only with the intersection of features measured in common across all studies, which results in the blind discarding of potentially useful feature information that is measured only in individual or subsets of all studies. We characterize the loss in predictive performance incurred by using only the intersection of feature information available across all studies when training predictors using gene expression data from microarray and sequencing datasets. We study the properties of linear and polynomial regression for imputing discarded features and demonstrate improvements in the external performance of predictors through simulation and in gene expression data collected on breast cancer patients. We propose and evaluate a pairwise imputation strategy that imputes cross-study missing features in each pair of studies and averages imputed features across pairs. Finally, we provide insights on which subsets of intersected and study-specific features should be used so that missing-feature imputation best promotes cross-study replicability. All code with directions to reproduce results in this paper is available at https://github.com/YujieWuu/Pairwise_imputation

bioinformatics↗

bioRxiv2021

Determinants of functional synaptic connectivity among amygdala-projecting prefrontal cortical neurons

The medial prefrontal cortex (mPFC) mediates a variety of complex cognitive functions via its vast and diverse connections with cortical and subcortical structures. Understanding the patterns of synaptic connectivity that comprise the mPFC local network is crucial for deciphering how this circuit processes information and relays it to downstream structures. To elucidate the synaptic organization of the mPFC, we developed a high-throughput optogenetic method for mapping large-scale functional synaptic connectivity. We show that mPFC neurons that project to the basolateral amygdala display unique spatial patterns of local-circuit synaptic connectivity within the mPFC, which distinguish them from the general mPFC cell population. Moreover, the intrinsic properties of the postsynaptic mPFC cell and anatomical position of both cells jointly account for ~7.5% of the variation in probability of connection between mPFC neurons, with anatomical distance and laminar position explaining most of this fraction in variation. Our findings demonstrate a functional segregation of mPFC excitatory neuron subnetworks, and reveal the factors determining connectivity in the mPFC.

neuroscience↗