Biology subjects

Samanta, S.

Publications and source records attributed to Samanta, S..

At least 19 recordsLinked to original sources

bioRxiv2026

Gasdermin E couples mitochondrial stress to STING-driven neuronal pyroptosis during Chandipura virus encephalitis

Neurotropic RNA viruses are major causes of fatal encephalitis worldwide, yet how infected neurons transition from antiviral defense to inflammatory cell death is not well characterized. Chandipura virus (CHPV), an emerging neurotropic rhabdovirus, causes acute, rapidly progressive encephalitis with high case fatality in children, but the mechanisms underlying its neuropathogenesis remain poorly defined. Here, we demonstrate that CHPV suppresses canonical RNA virus sensing early but subsequently switches to a mitochondria-driven innate immune program that culminates in inflammatory cell death. Early infection of neuronal cells with CHPV was marked by reduced levels of the mitochondrial antiviral adaptor protein, MAVS and attenuation of type I and III interferon responses. As infection progressed, mitochondrial dysfunction promoted accumulation of mtROS, mitochondrial accumulation of cleaved GSDME and cytosolic mtDNA release, triggering STING activation, which coincided with robust neuroinflammation and pyroptotic cell death. Pharmacological inhibition or genetic silencing of STING markedly attenuated inflammatory signaling, prevented pyroptotic membrane rupture and protected neurons from cell death without significantly affecting viral replication. In contrast, GSDME depletion reduced both viral replication and neuronal death. Notably, GSDME depletion markedly attenuated STING phosphorylation, while STING depletion also reduced GSDME activation, revealing functional coupling between these pathways during CHPV-induced neuronal injury. Collectively, our findings identify a mitochondria-GSDME-STING axis linking early immune evasion to neuroinflammation during CHPV infection, revealing a previously unrecognized mechanism of inflammatory neuronal death in viral encephalitis and highlighting STING as a potential therapeutic target in certain CNS viral infections.

immunology↗

bioRxiv2026

Deep interpretable learning of sample representations for characterizing disease states in single-cell transcriptomics

Single-cell transcriptomics technology offers unprecedented insights into molecular heterogeneity. However, capturing sample-level representations that reflect both systemic and cellular states remains challenging, especially when disease annotations are mostly available as coarse sample-level labels. Here, we introduce Phenoverse, an interpretable deep learning framework that learns sample-level disease state representations through cell type-aware residual encoding, prototype learning, and Perceiver-based aggregation. Applied to independent single-cell transcriptomic cohorts of COVID-19, Alzheimers disease, and systemic lupus erythematosus, totaling over 5 million cells, we demonstrate that learned sample representations enable disease state prediction and encode a continuous spectrum of disease severity on unseen data that correlate with multiple clinical and pathological measures, despite being trained solely on binary phenotype labels. Further, we demonstrate that trajectory-derived genes reveal cross-cohort molecular programs and show consistently higher reproducibility than traditional case-control comparisons. Finally, prototype learning provides intrinsic model interpretability and enables the characterization of cell type-specific disease states. Taken together, Phenoverse offers an interpretable disease-phenotyping approach to dissecting sample heterogeneity, and our results highlight its utility in translating complex single-cell transcriptomic data into patient-level biological insights.

bioinformatics↗

bioRxiv2026

Leishmania major targets macrophage Syntaxin-2 to impair phagolysosome biogenesis and promote intracellular survival

Macrophages destroy pathogens by engulfing them into phagosomes that mature into degradative phagolysosomes via lysosome fusion. Leishmania parasites subvert this antimicrobial pathway to establish intracellular infection and cause leishmaniasis. We previously identified the SNARE protein syntaxin-2 (Stx2) as a promoter of phagolysosome biogenesis that simultaneously limits particle binding and uptake. Consistent with this dual role, Stx2-depleted macrophages (Stx2-KD) show enhanced binding and internalization of Leishmania major. Stx2-KD macrophages also sustain higher intracellular parasite loads. We find that L. major actively targets macrophage Stx2 by selectively depleting Stx2 from phagosomes through its virulence metalloprotease GP63. Phagosomes containing GP63-deficient L. major retain Stx2 and acquire increased levels of lysosomal hydrolases and v-ATPase, restoring degradative capacity. In BALB/c mice, L. major infection markedly reduces Stx2 in infected tissues in a GP63-dependent manner. Collectively, our findings identify GP63-mediated Stx2 depletion as a key virulence strategy of L. major, positioning the GP63-Stx2 axis as a promising therapeutic target for leishmaniasis.

Cell Biology↗

bioRxiv2026

A multitask single-cell analysis framework with knowledge graph as a prior

Single-cell RNA sequencing has reshaped cancer biology by revealing the cellular and molecular diversity of the tumor microenvironment, yet most analysis pipelines still treat clustering, annotation, pathway scoring, metabolic inference, and cell-cell communication as separate tasks. Here, we present scKNIFE (Knowledge graph-based NMF for Inference of Functional Entities), a unified graph-regularized non-negative matrix factorization framework that jointly reconstructs single-cell expression profiles and infers activities of diverse biological entities, including pathways, cell-type markers, metabolic reactions and tasks, ligand-receptor interactions, and cell-state programs. By embedding a heterogeneous knowledge graph directly into the factorization objective, scKNIFE propagates information between biologically related entities through Laplacian regularization while sparsity constraints preserve interpretability. The integrated prior graph spans 47,274 nodes and 506,620 edges, assembled from complementary resources including Gene Ontology, Reactome, Hallmark gene sets, Human-GEM, CellChat, PanglaoDB, Cytopus, and Tabula Muris-derived annotations. Across multiple cancer-focused single-cell datasets, scKNIFE yields competitive to leading performance for clustering and cell-type annotation, while also recovering biologically coherent metabolic programs that agree with dedicated metabolism-focused methods and capture therapy-response-associated states. In addition, the framework supports downstream inference of cell-type-specific biological activities from a single latent representation. Together, these results establish scKNIFE as a modular and extensible framework for end-to-end biological interpretation of single-cell cancer transcriptomics

cancer biology↗

bioRxiv2026

Deterministic retrieval recovers biomedical associations lost by language models

Biomedical research increasingly relies on expert-curated databases to connect diseases, genes, variants, phenotypes, pathways and therapeutics. Incomplete or irreproducible retrieval can distort interpretation, mechanistic inference, variant assessment or therapeutic prioritization despite an unchanged evidence base. Agentic natural-language-to-SQL (NL2SQL) systems and Model Context Protocol-enabled agents can autonomously retrieve evidence from biomedical databases, including Open Targets and the Highly Confident Drug-Target Database, in response to natural-language questions. However, they may truncate large result sets, miss records when terminology differs from database vocabulary and return different outputs across runs. We systematically quantify these failures and introduce BioChirp, a goal-directed autonomous system for reliable retrieval from structured biomedical databases, built on interpretation-execution separation. A coordinated language-model layer interprets intent and resolves database-field assignments, after which entity resolution maps user terms to database terminology. A deterministic Steiner-tree planner and executor then retrieve matching records without further language-model involvement. Across curated biomedical databases spanning more than one million associations, BioChirp retrieved thousands of records for exhaustive queries, whereas baseline agents returned at most a few hundred or failed. Across 70 queries, five runs and three BioChirp database backends, BioChirp achieved a median cross-run Jaccard similarity of 1.0. Across 910 database-question evaluations derived from expert-written BioASQ questions and spanning ten databases, BioChirp retrieved at least one database record in 85.4% of cases and produced a fully correct answer in 51.1% of cases, compared with 44.6% and 17.4% for agentic NL2SQL. BioChirp is publicly available at https://biochirp.iiitd.edu.in.

bioinformatics↗

bioRxiv2026

A Network-Based Approach for Prioritizing Candidate Genes in Alzheimer's Disease

Alzheimers disease (AD) is a multifactorial neurodegenerative disorder characterized by coordinated dysregulation of multiple genes, requiring system-level approaches to elucidate underlying molecular mechanisms. While existing computational studies largely focus on differential expression analysis or machine-learning-based feature selection, they often overlook inter-gene relationships and network topology, limiting biological interpretability. In this study, we present a network-based framework for prioritizing candidate genes in Alzheimers disease by integrating gene co-expression network analysis with multiple centrality measures. Transcriptomic data comprising approximately 39,000 genes across 324 Alzheimers and control samples were preprocessed using log-transformation, variance filtering and Z-score normalization, followed by LASSO-based feature selection to retain phenotype-associated genes. A weighted gene co-expression network was constructed using Pearson correlation to capture coordinated transcriptional activity. Network topology was analyzed using degree, betweenness and eigenvector centrality to identify genes that are highly connected, act as information brokers or occupy influential positions within the network. A consensus ranking was derived by merging these centrality measures, enabling robust prioritization of candidate genes. The results highlight a subset of highly central genes, including several small nucleolar RNAs and regulatory transcripts implicated in RNA processing, synaptic function and neurodegenerative pathways. By jointly leveraging co-expression structure and complementary centrality metrics, the proposed framework provides an interpretable and reproducible strategy for identifying biologically meaningful gene candidates, offering potential utility for biomarker discovery and therapeutic target prioritization in Alzheimers disease.

systems biology↗

bioRxiv2026

AADAT-Driven Metabolic Control of Malate and CoQ10 Shapes Immune Evasion in Triple-Negative Breast Cancer

Compared to other subtypes of breast cancer, triple-negative breast cancers (TNBC) have fewer treatment options and exhibit a worse prognosis. Through integrated transcriptomic, metabolomic, immunohistochemical, spatial, and clinical analyses, we identify the mitochondrial enzyme, -aminoadipate aminotransferase (AADAT) as a previously unrecognized metabolic immune checkpoint in TNBC. AADAT mRNA and protein were significantly upregulated in human TNBC, and high AADAT expression was associated with reduced intra-tumoral CD8 T-cell density and inferior survival. Genetic silencing of AADAT in orthotopic murine TNBC models curtailed primary tumor growth and distant metastasis in a CD8 T-cell-dependent manner, enhanced effector T-cell activation, and sensitized tumors to dual PD-1/CTLA-4 blockade. Mechanistically, unbiased metabolomics showed increased malate levels after AADAT knockdown. Additionally, 4-hydroxyphenylpyruvate, an essential precursor for coenzyme Q10(CoQ10) biosynthesis, decreased following AADAT knockdown, suggesting an impaired mitochondrial electron transport chain. CoQ10 supplementation restored metabolic balance and reversed malate accumulation caused by AADAT knockdown, indicating that AADAT helps maintain CoQ10-supported redox homeostasis, thereby preventing malate buildup and export. Notably, malate addition directly boosted CD8 T-cell oxidative metabolism, increased the NAD/NADH ratio and reactive oxygen species, and augmented TNF- and IFN-{gamma} production. In vivo, malate supplementation in drinking water phenocopied AADAT knockdown, restored the response to paclitaxel plus anti-PD-1 therapy in multiple independent syngeneic TNBC models with de novo or acquired resistance to immunotherapy, reduced tumor burden, and prolonged survival. In patient cohorts, higher spatially clustered intra-tumoral malate is associated with co-localization of functional CD8 T cells, decreased exhausted T-cell neighborhoods, and superior post-chemotherapy outcomes. These data position AADAT as a central metabolic orchestrator of immune escape in TNBC and nominate oral malate as a readily translatable adjuvant to reverse chemo-immunotherapy resistance in TNBC. Statement of SignificanceAADAT defines a metabolic-immune axis driving immune evasion and therapy resistance in triple-negative breast cancer. Blocking AADAT or administering oral malate reactivates CD8 T-cell immunity and sensitizes chemo-immunotherapy-resistant tumors to these agents. These findings uncover a readily translatable metabolic vulnerability with potential to improve outcomes for patients with aggressive breast cancer subtypes.

cancer biology↗

bioRxiv2026

An Indole Dimer Antifungal Metabolite from a Rice Seed Endophyte Inhibits Ergosterol Biosynthesis in Fungal Pathogens

The increasing prevalence of fungal phytopathogens and the widespread emergence of fungicide resistance necessitate alternative antifungal strategies with reduced environmental impact. Here, we report the isolation and characterization of a novel antifungal metabolite, SM06, produced by the rice seed-associated endophytic bacterium Phytobacter sp. RSE02. SM06 exhibited broad-spectrum antifungal activity against plant and human pathogenic fungi, including Curvularia lunata, Fusarium oxysporum, and Candida albicans. In vitro assays and micromorphological analyses revealed that SM06, an indole dimer, disrupts fungal cell membrane integrity, while in planta experiments demonstrated significant suppression of brown leaf spot disease in tomato and rice. Molecular docking suggested that SM06 binds to lanosterol 14-demethylase (ERG11), a key enzyme in fungal sterol biosynthesis. Consistent with this prediction, LC-MS-based analyses confirmed a significant reduction in ergosterol content in SM06-treated fungal cells. Together, these findings identify SM06 as a biologically active antifungal metabolite produced by a plant-associated bacterium and highlight its potential application in sustainable fungal disease management. IMPORTANCEFungal diseases cause major losses in crop-production and contribute to the growing challenge of antifungal resistance, underscoring the need for sustainable alternatives to chemical fungicides. This study identifies SM06, a novel indole dimer produced by the rice seed endophyte Phytobacter sp. RSE02, with strong antifungal activity against economically important plant pathogens and clinically relevant fungi. Through integrated chemical, cellular, and in planta analyses, we demonstrate that SM06 disrupts fungal membrane integrity by inhibiting ergosterol biosynthesis. The compound is biocompatible, stable, and effective in plant disease suppression, highlighting its translational potential for crop protection. These findings reveal seed endophytes as an important yet underexplored sources of antifungal metabolites and provide a mechanistic foundation for developing eco-friendly biocontrol strategies with implications beyond agriculture. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=179 SRC="FIGDIR/small/697688v1_ufig1.gif" ALT="Figure 1"> View larger version (86K): org.highwire.dtl.DTLVardef@c7da68org.highwire.dtl.DTLVardef@d3f1a1org.highwire.dtl.DTLVardef@9cab82org.highwire.dtl.DTLVardef@1c01141_HPS_FORMAT_FIGEXP M_FIG C_FIG

microbiology↗

bioRxiv2025

Leishmania secretory factor GP63 targets the DICER1/miR-122/hepcidin axis in host macrophages to deplete Nramp1

Micronutrient sequestration is a powerful host defence mechanism against intracellular pathogens. Central to this is the iron transporter Nramp1, which actively effluxes iron from phagolysosomes, depriving the engulfed pathogens of this essential element. Leishmania have evolved a clever strategy to counteract this by triggering hepcidin-mediated proteasomal degradation of Nramp1. Interestingly, treatment with Leishmania major conditioned media induced hepcidin upregulation and Nramp1 degradation even in uninfected macrophages, with simultaneous increase in endo/lysosomal iron content. This suggested the involvement of a parasite-derived secretory factor, ultimately leading to the identification of the metalloprotease GP63 as the effector responsible for Nramp1 degradation. Conditioned medium from the GP63 knockout strain (LmGP63-/-) failed to upregulate hepcidin expression or promote Nramp1 degradation. Further, we show that GP63 depletes DICER1 in macrophages, blocking maturation of miR-122, a known negative regulator of hepcidin. These in vitro findings were validated in a mouse model, where the wild type L. major, but not the LmGP63-/- strain, could deplete DICER1, induce hepcidin expression and reduce Nramp1 levels in the infected tissue. Collectively, we uncover a novel role for Leishmania GP63 in targeting the DICER1/miR-122 axis in host macrophages to trigger hepcidin expression and Nramp1 degradation, thereby facilitating iron acquisition by the parasite.

cell biology↗

bioRxiv2024

Syntaxin-2 inhibits phagocytic uptake but promotes phagolysosome formation in macrophages

Phagocytosis engulfs receptor-bound particles by structuring phagocytic cups that expand to form phagosome vesicles. Phagosomes fuse with endosomes and lysosomes to gradually mature into acidic, hydrolase-enriched phagolysosomes for content degradation. While an essential cellular process for eliminating unwanted substances to secure host defense and organismal homeostasis, defective or uncontrolled phagocytosis can be detrimental. Here, we report syntaxin-2 (Stx2), a poorly characterized SNARE in phagocytes, defines the course of macrophage phagocytosis by coordinating cell surface receptor density, phagocytic cup expansion, and phagosome maturation. Stx2 is expressed on the plasma membrane, phagosomes, early endosomes, and to some extent on recycling endosomes, late endosomes and VAMP4-positive compartments. Stx2 knockdown (Stx2-KD) augments engagement and uptake of IgG-opsonized particles by enhanced surface expression of Fc receptors (FcR) and efficient expansion of phagocytic cups. These events are driven by increased recycling of FcR, and enhanced delivery of early endosomes and VAMP4-positive post-Golgi compartments to phagocytic cups. Interestingly, Stx2-KD macrophages exhibit reduced secretion of pro-cathepsins with concomitant increase in lysosome content. However, Stx2 depletion prevents phagosome coalescence with late endosomes and lysosomes to impair maturation by depleted acquisition of cathepsins and reduced acidification. Consequently, macrophages depleted of Stx2 manifest uncontrolled uptake of IgG-opsonized Escherichia coli and impaired digestion resulting in increased bacterial load. Collectively, we uncovered a previously unknown role of Stx2 as a critical balancer of phagocytic uptake and phagolysosomal clearance in macrophages, suggesting that it could be an attractive target for modulation of phagocytosis plasticity and to control aberrant phagocytosis.

cell biology↗

bioRxiv2024

Binding profiles for 954 Drosophila and C. elegans transcription factors revealtissue specific regulatory relationships

A catalog of transcription factor (TF) binding sites in the genome is critical for deciphering regulatory relationships. Here we present the culmination of the modERN (model organism Encyclopedia of Regulatory Networks) consortium that systematically assayed TF binding events in vivo in two major model organisms, Drosophila melanogaster (fly) and Caenorhabditis elegans (worm). We describe key features of these datasets, comprising 604 TFs identifying 3.6M sites in the fly and 350 TFs identifying 0.9 M sites in the worm. Applying a machine learning model to these data identifies sets of TFs with a prominent role in promoting target gene expression in specific cell types. TF binding data are available through the ENCODE Data Coordinating Center and at https://epic.gs.washington.edu/modERNresource, which provides access to processed and summary data, as well as widgets to probe cell type-specific TF-target relationships. These data are a rich resource that should fuel investigations into TF function during development.

genomics↗

bioRxiv2024

Functional diversity in the Hsp60 of Sulfolobus acidocaldarius: mosaic of Group I and Group II chaperonin

External stress can disrupt protein homeostasis in organisms, necessitating the involvement of heat shock proteins (Hsps) to restore equilibrium and ensure survival. Unlike other organisms, the thermoacidophilic crenarchaeon Sulfolobus acidocaldarius lacks Hsp100, Hsp90, and Hsp70, possessing only two small heat shock proteins (Hsp14 and Hsp20) and one group II chaperonin, Hsp60. This raises questions about how protein substrates are protected and transferred to Hsp60 for refolding without other chaperones. Our study focused on ATP-dependent Hsp60 in S. acidocaldarius, revealing its formation of oligomeric structures in the presence of ATP. While ATP hydrolysis is not essential for oligomer formation and lid closure, it is crucial for Hsp60s chaperone activity, effectively folding stress-denatured substrate proteins by stabilizing their folded conformations. The mechanism involves hydrophobic recognition of unfolded substrates, encapsulating and releasing them in a more folded state. Negatively charged inner surface of the ring seems to be responsible for driving the folding of the substrate. Importantly, Hsp14 was found to transfer substrate proteins to Hsp60{beta}, orchestrating their refolding into an active state. Beyond protein folding, Hsp60{beta} protects the membrane under stress, contributing to maintaining membrane rigidity. Hsp60 exhibits nested cooperativity in ATPase activity, adapting to ATP concentration changes and interestingly Hsp60{beta} and Hsp60{beta} complex shows a mosaic behaviour during ATP hydrolysis belonging to both group I and group II chaperonin respectively. In conclusion, our study provides insights into the intricate mechanisms employed by Hsp60 in S. acidocaldarius to maintain protein homeostasis. It offers a comprehensive understanding of Hsp60s role in the heat shock response pathway, shedding light on fundamental cellular processes in extremophilic archaea.

biochemistry↗

bioRxiv2023

Leishmania major-induced alteration of host cellular and systemic copper homeostasis drives the fate of infection

Copper plays a key role in host-pathogen interaction. We found that during Leishmania major infection, the parasite-harboring macrophage regulates its copper homeostasis pathway in a way to facilitate copper-mediated neutralization of the pathogen. Copper-ATPase ATP7A transports copper to amastigote-harboring phagolysosomes to induce stress on parasites. Leishmania in order to evade the copper stress, utilizes a variety of manipulative measures to lower the host-induced copper stress. It induces deglycosylation and degradation of host-ATP7A and downregulation of copper importer, CTR1 by cysteine oxidation. Additionally, Leishmania induces CTR1 endocytosis that arrests copper uptake. In mouse model of infection, we report an increase in systemic bioavailable copper in infected animals. Heart acts as the major organ for diverting its copper reserves to systemic circulation to fight-off infection by downregulating its CTR1. Our study explores reciprocal mechanism of manipulation of host copper homeostasis pathway by macrophage and Leishmania to gain respective advantages in host-pathogen interaction.

cell biology↗

bioRxiv2023

A novel role of macrophage PIST protein in regulating Leishmania major infection

PDZ protein interacting specifically with Tc10 or PIST is a mammalian trans-Golgi resident protein that regulates subcellular sorting of plasma membrane receptors. PIST has recently been found to play an important role in regulating viral pathogenesis. Alteration in PIST expression is linked to the reprogramming of cell surface receptors which is crucial in determining herpes simplex virus1 infection. In this context, PIST is crucial in triggering autophagy via Beclin 1 -PI3KC3 pathway. However, there is complete lack in our knowledge on the role of this protein in any parasitic infection. Leishmania parasites infect their host macrophage cells via phagocytosis and their survival within the parasitophorous compartment has recently been found to be dependent on host autophagy by a yet to be identified mechanism. Using Leishmania major (L. major)-macrophage infection model system we, for the first-time report here that in infected macrophages Golgi resident PIST protein migrates towards the parasite containing compartment. We have also found that PIST associates with Beclin 1, however, not with LC3 within L. major parasite containing compartment of infected macrophages. Further, we performed genetic ablation of PIST by siRNA and observed that knockdown of macrophage PIST in turn helps in parasite replication. Contrary to this, overexpression of PIST in macrophages restricted the multiplication of L. major. Collectively, our study for the first time reveals that PIST is essential in regulating intracellular parasite, L. major infection within macrophage cells. Summary StatementMammalian PIST protein plays a crucial role in regulating cellular trafficking events. Here, we showed that PIST status is altered within Leishmania major parasite infected macrophages. Further, we confirmed that PIST is essential in restricting parasite growth. Additionally, a potential interacting axis between PIST and Beclin 1 is identified during infection.

cell biology↗

bioRxiv2022

Heparin-derived theranostic nanoprobes overcome the blood-brain barrier and target glioma in murine model

The poor permeability of theranostic agents across the blood-brain-barrier (BBB) significantly hampers the development of new treatment modalities for neurological diseases. We have discovered a new biomimetic nanocarrier using heparin (HP) that effectively passes the BBB and targets glioblastoma. Specifically, we designed HP coated gold nanoparticles (HP-AuNPs) that were labeled with three different imaging modalities namely, fluorescein (FITC-HP-AuNP), radioisotope 68Gallium (68Ga-HP-AuNPs), and MRI active gadolinium (Gd-HP-AuNPs). The systemic infusion of FITC-HP-AuNPs in three different mouse strains (C57BL/6JRj, FVB, and NMRI-nude) displayed excellent penetration and revealed uniform distribution of fluorescent particles in the brain parenchyma (69-86%) with some accumulation in neurons (8-18%) and microglia (4-10%). Tail-vein administration of radiolabeled 68Ga-HP-AuNPs in healthy rats also showed 68Ga-HP-AuNP inside the brain parenchyma and in areas containing cerebrospinal fluid, such as the lateral ventricles, the cerebellum, and brain stem. Finally, tail-vein administration of Gd-HP-AuNPs (that display [~]3 fold higher relaxivity than that of commercial Gd-DTPA) in an orthotopic glioblastoma (U87MG xenograft) model in nude mice demonstrated enrichment of T1-contrast at the intracranial tumor with a gradual increase in the contrast in the tumor region between 1h-3h. We believe, our finding offers the untapped potential of HP-derived-NPs to deliver cargo molecules for treating neurological disorders.

bioengineering↗

bioRxiv2021

MassGenie: a transformer-based deep learning method for identifying small molecules from their mass spectra

The inverse problem of mass spectrometric molecular identification ( given a mass spectrum, calculate the molecule whence it came) is largely unsolved, and is especially acute in metabolomics where many small molecules remain unidentified. This is largely because the number of experimentally available electrospray mass spectra of small molecules is quite limited. However, the forward problem ( calculate a small molecules likely fragmentation and hence at least some of its mass spectrum from its structure alone) is much more tractable, because the strengths of different chemical bonds are roughly known. This kind of molecular identification problem may be cast as a language translation problem in which the source language is a list of high-resolution mass spectral peaks and the translation a representation (for instance in SMILES) of the molecule. It is thus suitable for attack using the deep neural networks known as transformers. We here present MassGenie, a method that uses a transformer-based deep neural network, trained on ~6 million chemical structures with augmented SMILES encoding and their paired molecular fragments as generated in silico, explicitly including the protonated molecular ion. This architecture (containing some 400 million elements) is used to predict the structure of a molecule from the various fragments that may be expected to be observed when some of its bonds are broken. Despite being given essentially no detailed nor explicit rules about molecular fragmentation methods, isotope patterns, rearrangements, neutral losses, and the like, MassGenie learns the effective properties of the mass spectral fragment and valency space, and can generate candidate molecular structures that are very close or identical to those of the true molecules. We also use VAE-Sim, a previously published variational autoencoder, to generate candidate molecules that are similar to the top hit. In addition to using the top hits directly, we can produce a rank order of these by round-tripping candidate molecules and comparing them with the true molecules, where known. As a proof of principle, we confine ourselves to positive electrospray mass spectra from molecules with a molecular mass of 500Da or lower. The transformer method, applied here for the first time to mass spectral interpretation, works extremely effectively both for mass spectra generated in silico and on experimentally obtained mass spectra from pure compounds. The ability to create and to learn millions of fragmentation patterns in silico, and therefrom generate candidate structures (that do not have to be in existing libraries) directly, thus opens up entirely the field of de novo small molecule structure prediction from experimental mass spectra.

bioinformatics↗

bioRxiv2020

Naphthalene monoimide derivative ameliorates amyloid burden and cognitive decline in a transgenic mouse model of Alzheimer's disease

Alzheimers disease (AD) is a major neurodegenerative disorder and the leading cause of dementia worldwide. Predominantly, misfolding and aggregation of amyloid-{beta} (A{beta}) peptides associated with multifaceted toxicity is the neuropathological hallmark of AD pathogenesis and thus, primary therapeutic target to ameliorate neuronal toxicity and cognitive deficits. Herein, we report the design, synthesis and evaluation of small molecule inhibitors with naphthalene monoimide scaffold to ameliorate in vitro and in vivo amyloid induced neurotoxicity. The detailed studies established TGR63 as the lead candidate to rescue neuronal cells from amyloid toxicity. The in silico studies showed disruption of salt bridges and intermolecular hydrogen bonding interactions within A{beta}42 fibrils by the interaction of TGR63, causing destabilization of A{beta}42 assembly. Remarkably, TGR63 treatment showed a significant reduction in cortical and hippocampal amyloid burden in the progressive stages of APP/PS1 AD mice brain. Various behavioral tests demonstrated rescued cognitive deficits. The excellent biocompatibility, BBB permeability and therapeutic efficacy to reduce amyloid burden make TGR63 a promising candidate for the treatment of AD.

neuroscience↗

bioRxiv2020

VAE-Sim: a novel molecular similarity measure based on a variational autoencoder

Molecular similarity is an elusive but core unsupervised cheminformatics concept, yet different fingerprint encodings of molecular structures return very different similarity values even when using the same similarity metric. Each encoding may be of value when applied to other problems with objective or target functions, implying that a priori none is better than the others, nor than encoding-free metrics such as maximum common substructure (MCSS). We here introduce a novel approach to molecular similarity, in the form of a variational autoencoder (VAE). This learns the joint distribution p(z|x) where z is a latent vector and x are the (same) input/output data. It takes the form of a bowtie-shaped artificial neural network. In the middle is a bottleneck layer or latent vector in which inputs are transformed into, and represented as, a vector of numbers (encoding), with a reverse process (decoding) seeking to return the SMILES string that was the input. We train a VAE on over 6 million druglike molecules and natural products (including over one million in the final holdout set). The VAE vector distances provide a rapid and novel metric for molecular similarity that is both easily and rapidly calculated. We describe the method and its application to a typical similarity problem in cheminformatics.

pharmacology and toxicology↗