Biomedicines. 2024 Jul 22;12(7):1625. doi: 10.3390/biomedicines12071625.

ABSTRACT

(1) Background: The rise in antibiotic resistant bacteria poses a significant threat to public health worldwide, necessitating innovative solutions. This study explores the role of Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) in the context of antibiotic resistance among different species from the Enterococcus genus. (2) Methods: The genomes of Enterococcus included in the study were analyzed using CRISPRCasFinder to distinguish between CRISPR-positive (level 4 CRISPR) and CRISPR-negative genomes. Antibiotic resistance genes were identified, and a comparative analysis explored potential associations between CRISPR presence and antibiotic resistance profiles in Enterococcus species. (3) Results: Out of ten antibiotic resistance genes found in Enterococcus species, only one, the efmA gene, showed a strong association with CRISPR-negative isolates, while the others did not significantly differ between CRISPR-positive and CRISPR-negative Enterococcus genomes. (4) Conclusion: These findings indicate that the efmA gene may be more prevalent in CRISPR-negative Enterococcus genomes, and they may contribute to a better understanding of the molecular mechanisms underlying the acquisition of antibiotic resistance genes in Enterococcus species.

PMID:39062198 | DOI:10.3390/biomedicines12071625

Biomedicines. 2024 Jul 5;12(7):1496. doi: 10.3390/biomedicines12071496.

ABSTRACT

The field of multi-omics has witnessed unprecedented growth, converging multiple scientific disciplines and technological advances. This surge is evidenced by a more than doubling in multi-omics scientific publications within just two years (2022-2023) since its first referenced mention in 2002, as indexed by the National Library of Medicine. This emerging field has demonstrated its capability to provide comprehensive insights into complex biological systems, representing a transformative force in health diagnostics and therapeutic strategies. However, several challenges are evident when merging varied omics data sets and methodologies, interpreting vast data dimensions, streamlining longitudinal sampling and analysis, and addressing the ethical implications of managing sensitive health information. This review evaluates these challenges while spotlighting pivotal milestones: the development of targeted sampling methods, the use of artificial intelligence in formulating health indices, the integration of sophisticated n-of-1 statistical models such as digital twins, and the incorporation of blockchain technology for heightened data security. For multi-omics to truly revolutionize healthcare, it demands rigorous validation, tangible real-world applications, and smooth integration into existing healthcare infrastructures. It is imperative to address ethical dilemmas, paving the way for the realization of a future steered by omics-informed personalized medicine.

PMID:39062068 | DOI:10.3390/biomedicines12071496

Antioxidants (Basel). 2024 Jun 27;13(7):782. doi: 10.3390/antiox13070782.

ABSTRACT

Cadmium (Cd) toxicity poses a significant threat to human health and the environment due to its widespread occurrence and persistence. In recent years, considerable attention has been directed towards exploring natural compounds with potential protective effects against Cd-induced toxicity. Among these compounds, resveratrol (RV) has emerged as a promising candidate, demonstrating a range of beneficial effects attributed to its antioxidant and anti-inflammatory properties. This literature review systematically evaluates the protective role of RV against Cd toxicity, considering the various mechanisms of action involved. A comprehensive analysis of both in vitro and in vivo studies is conducted to provide a comprehensive understanding of RV efficacy in mitigating Cd-induced damage. Additionally, this review highlights the importance of phytoremediation strategies in addressing Cd contamination, emphasizing the potential of RV in enhancing the efficiency of such remediation techniques. Through the integration of diverse research findings, this review underscores the therapeutic potential of RV in combating Cd toxicity and underscores the need for further investigation to elucidate its precise mechanisms of action and optimize its application in environmental and clinical settings.

PMID:39061851 | DOI:10.3390/antiox13070782

Cell Commun Signal. 2024 Jul 26;22(1):377. doi: 10.1186/s12964-024-01742-3.

ABSTRACT

BACKGROUND: Tumor heterogeneity is a main contributor of resistance to anti-cancer targeted agents though it has proven difficult to study. Unfortunately, model systems to functionally characterize and mechanistically study dynamic responses to treatment across coexisting subpopulations of cancer cells remain a missing need in oncology.

METHODS: Using single cell cloning and expansion techniques, we established monoclonal cell subpopulations (MCPs) from a commercially available epidermal growth factor receptor (EGFR)-mutant non-small cell lung cancer cell line. We then used this model sensitivity to the EGFR inhibitor osimertinib across coexisting cell populations within the same tumor. Pathway-centered signaling dynamics associated with response to treatment and morphological characteristics of the MCPs were assessed using Reverse Phase Protein Microarray. Signaling nodes differentially activated in MCPs less sensitive to treatment were then pharmacologically inhibited to identify target signaling proteins putatively implicated in promoting drug resistance.

RESULTS: MCPs demonstrated highly heterogeneous sensitivities to osimertinib. Cell viability after treatment increased > 20% compared to the parental line in selected MCPs, whereas viability decreased by 75% in other MCPs. Reduced treatment response was detected in MCPs with higher proliferation rates, EGFR L858R expression, activation of EGFR binding partners and downstream signaling molecules, and expression of epithelial-to-mesenchymal transition markers. Levels of activation of EGFR binding partners and MCPs' proliferation rates were also associated with response to c-MET and IGFR inhibitors.

CONCLUSIONS: MCPs represent a suitable model system to characterize heterogeneous biomolecular behaviors in preclinical studies and identify and functionally test biological mechanisms associated with resistance to targeted therapeutics.

PMID:39061010 | DOI:10.1186/s12964-024-01742-3

Nat Struct Mol Biol. 2024 Jul 26. doi: 10.1038/s41594-024-01367-7. Online ahead of print.

ABSTRACT

The biogenesis and maintenance of thylakoid membranes require vesicle-inducing protein in plastids 1 (VIPP1). VIPP1 is a member of the endosomal sorting complex required for transport-III (ESCRT-III) superfamily, whose members form diverse filament-based supramolecular structures that facilitate membrane deformation and fission. VIPP1 cryo-electron microscopy (EM) structures in solution revealed helical rods and baskets of stacked rings, with amphipathic membrane-binding domains in the lumen. However, how VIPP1 interacts with membranes remains largely unknown. Here, using high-speed atomic force microscopy (HS-AFM), we show that VIPP1 assembles into right-handed chiral spirals and regular polygons on supported lipid bilayers via ESCRT-III-like filament assembly and dynamics. VIPP1 filaments grow clockwise into spirals through polymerization at a ring-shaped central polymerization hub, and into polygons through clockwise polymerization at the sector peripheries. Interestingly, VIPP1 initially forms Archimedean spirals, which upon maturation transform into logarithmic spirals through lateral annealing of strands to the outermore low-curvature spiral turns.

PMID:39060677 | DOI:10.1038/s41594-024-01367-7

Nat Plants. 2024 Jul 26. doi: 10.1038/s41477-024-01760-6. Online ahead of print.

ABSTRACT

In most plant species, sepals-the outermost floral organs-provide a protective shield for reproductive organs. How the floral bud becomes sealed is unknown. In Arabidopsis, we identified a small region at the sepal tip that is markedly curved inward early on and remains curved even after anthesis. Through modelling and quantitative growth analysis, we find that this hook emerges from growth arrest at the tip at a stage when cortical microtubules align with growth-derived tensile stress. Depolymerizing microtubules specifically at young sepal tips hindered hook formation and resulted in open floral buds. Mutants with defective growth pattern at the tip failed to curve inwards, whereas mutants with enhanced alignment of cortical microtubules at the tip exhibited a stronger hook. We propose that floral buds are locked due to a stress-derived growth arrest event curving the sepal tip and forming a rigid hook early on during flower development.

PMID:39060423 | DOI:10.1038/s41477-024-01760-6

Nat Commun. 2024 Jul 25;15(1):6279. doi: 10.1038/s41467-024-50393-y.

ABSTRACT

The molecular mechanisms by which FoxO transcription factors mediate diametrically opposite cellular responses, namely death and survival, remain unknown. Here we show that Mst1 phosphorylates FoxO1 Ser209/Ser215/Ser218/Thr228/Ser232/Ser243, thereby inhibiting FoxO1-mediated transcription of proapoptotic genes. On the other hand, Mst1 increases FoxO1-C/EBP-β interaction and activates C/EBP-β by phosphorylating it at Thr299, thereby promoting transcription of prosurvival genes. Myocardial ischemia/reperfusion injury is larger in cardiac-specific FoxO1 knockout mice than in control mice. However, the concurrent presence of a C/EBP-β T299E phospho-mimetic mutation reduces infarct size in cardiac-specific FoxO1 knockout mice. The C/EBP-β phospho-mimetic mutant exhibits greater binding to the promoter of prosurvival genes than wild type C/EBP-β. In conclusion, phosphorylation of FoxO1 by Mst1 inhibits binding of FoxO1 to pro-apoptotic gene promoters but enhances its binding to C/EBP-β, phosphorylation of C/EBP-β, and transcription of prosurvival genes, which stimulate protective mechanisms in the heart.

PMID:39060225 | DOI:10.1038/s41467-024-50393-y

J Biol Chem. 2024 Jul 24:107602. doi: 10.1016/j.jbc.2024.107602. Online ahead of print.

ABSTRACT

Glycosylation is a predominant strategy plants employ to fine-tune the properties of small molecule metabolites to affect their bioactivity, transport, and storage. It is also important in biotechnology and medicine as many glycosides are utilized in human health. Small molecule glycosylation is largely carried out by family 1 glycosyltransferases. Here, we report a structural and biochemical investigation of UGT95A1, a family 1 GT enzyme from Pilosella officinarum that exhibits a strong, unusual regiospecificity for the 3'-O position of flavonoid acceptor substrate luteolin. We obtained an apo crystal structure to help drive the analyses of a series of binding site mutants, revealing that while most residues are tolerant to mutations, key residues M145 and D464 are important for overall glycosylation activity. Interestingly, E347 is crucial for maintaining the strong preference for 3'-O glycosylation, while R462 can be mutated to increase regioselectivity. The structural determinants of regioselectivity were further confirmed in homologous enzymes. Our study also suggests that the enzyme contains large, highly dynamic, disordered regions. We showed that while most disordered regions of the protein have little to no implication in catalysis, the disordered regions conserved among investigated homologues are important to both the overall efficiency and regiospecificity of the enzyme. This report represents a comprehensive in-depth analysis of a family 1 GT enzyme with a unique substrate regiospecificity and may provide a basis for enzyme functional prediction and engineering.

PMID:39059496 | DOI:10.1016/j.jbc.2024.107602

Sci Adv. 2024 Jul 26;10(30):eado5716. doi: 10.1126/sciadv.ado5716. Epub 2024 Jul 26.

ABSTRACT

The three-dimensional (3D) organization of chromatin within the nucleus is crucial for gene regulation. However, the 3D architectural features that coordinate the activation of an entire chromosome remain largely unknown. We introduce an omics method, RNA-associated chromatin DNA-DNA interactions, that integrates RNA polymerase II (RNAPII)-mediated regulome with stochastic optical reconstruction microscopy to investigate the landscape of noncoding RNA roX2-associated chromatin topology for gene equalization to achieve dosage compensation. Our findings reveal that roX2 anchors to the target gene transcription end sites (TESs) and spreads in a distinctive boot-shaped configuration, promoting a more open chromatin state for hyperactivation. Furthermore, roX2 arches TES to transcription start sites to enhance transcriptional loops, potentially facilitating RNAPII convoying and connecting proximal promoter-promoter transcriptional hubs for synergistic gene regulation. These TESs cluster as roX2 compartments, surrounded by inactive domains for coactivation of multiple genes within the roX2 territory. In addition, roX2 structures gradually form and scaffold for stepwise coactivation in dosage compensation.

PMID:39058769 | DOI:10.1126/sciadv.ado5716

Sci Immunol. 2024 Jul 26;9(97):eadk3981. doi: 10.1126/sciimmunol.adk3981. Epub 2024 Jul 26.

ABSTRACT

Tumor-associated macrophages (TAMs) are a heterogeneous population of cells whose phenotypes and functions are shaped by factors that are incompletely understood. Herein, we asked when and where TAMs arise from blood monocytes and how they evolve during tumor development. We initiated pancreatic ductal adenocarcinoma (PDAC) in inducible monocyte fate-mapping mice and combined single-cell transcriptomics and high-dimensional flow cytometry to profile the monocyte-to-TAM transition. We revealed that monocytes differentiate first into a transient intermediate population of TAMs that generates two longer-lived lineages of terminally differentiated TAMs with distinct gene expression profiles, phenotypes, and intratumoral localization. Transcriptome datasets and tumor samples from patients with PDAC evidenced parallel TAM populations in humans and their prognostic associations. These insights will support the design of new therapeutic strategies targeting TAMs in PDAC.

PMID:39058763 | DOI:10.1126/sciimmunol.adk3981

J Exp Bot. 2024 Jul 26:erae326. doi: 10.1093/jxb/erae326. Online ahead of print.

ABSTRACT

GLABRA2 (GL2), a class IV homeodomain leucine-zipper (HD-Zip IV) transcription factor (TF) from Arabidopsis, is a developmental regulator of specialized cell types in the epidermis. GL2 contains a monopartite nuclear localization sequence (NLS) that is conserved in most HD-Zip IV members across the plants. We demonstrate that NLS mutations affect nuclear transport and result in a loss-of-function phenotypes. NLS fusions to EYFP show that it is sufficient for nuclear localization in roots and trichomes. Despite partial overlap of the NLS with the homeodomain, genetic dissection indicates that nuclear localization and DNA binding are separable functions. Affinity purification of GL2 from plants followed by MS-based proteomics identified Importin α (IMPα) isoforms as potential GL2 interactors. NLS structural prediction and molecular docking studies with IMPα-3 revealed major interacting residues. Cytosolic yeast two-hybrid assays and co-immunoprecipitation experiments with recombinant proteins verified NLS-dependent interactions between GL2 and several IMPα isoforms. IMPα triple mutants (impα-1,2,3) exhibit abnormal trichome formation and defects in GL2 nuclear localization in trichomes, consistent tissue-specific and redundant functions of IMPα isoforms. Taken together, our findings provide mechanistic evidence for IMPα-dependent nuclear localization of GL2 in Arabidopsis, a process that is critical for cell-type differentiation of the epidermis.

PMID:39058342 | DOI:10.1093/jxb/erae326

Metabolites. 2024 Jul 7;14(7):379. doi: 10.3390/metabo14070379.

ABSTRACT

Over the millennia, patterns of food consumption have changed; however, foods were always whole foods. Ultra-processed foods (UPFs) have been a very recent development and have become the primary food source for many people. The purpose of this review is to propose the hypothesis that, forsaking the evolutionary dietary environment, and its complex milieu of compounds resulting in an extensive metabolome, contributes to chronic disease in modern humans. This evolutionary metabolome may have contributed to the success of early hominins. This hypothesis is based on the following assumptions: (1) whole foods promote health, (2) essential nutrients cannot explain all the benefits of whole foods, (3) UPFs are much lower in phytonutrients and other compounds compared to whole foods, and (4) evolutionary diets contributed to a more diverse metabolome. Evidence will be presented to support this hypothesis. Nutrition is a matter of systems biology, and investigating the evolutionary metabolome, as compared to the metabolome of modern humans, will help elucidate the hidden connections between diet and health. The effect of the diet on the metabolome may also help shape future dietary guidelines, and help define healthy foods.

PMID:39057702 | DOI:10.3390/metabo14070379

Metabolites. 2024 Jun 28;14(7):368. doi: 10.3390/metabo14070368.

ABSTRACT

There is a growing interest in the discovery of novel xanthine oxidase inhibitors for gout prevention and treatment with fewer side effects. This study aimed to identify the xanthine oxidase (XO) inhibitory potential and drug-likeness of the metabolites present in the methanolic leaf extract of Anastatica (A.) hierochuntica L. using in vitro and in silico models. The extract-derived metabolites were identified by liquid-chromatography-quadrupole-time-of-flight-mass-spectrometry (LC-QTOF-MS). Molecular docking predicted the XO inhibitory activity of the identified metabolites and validated the best scored in vitro XO inhibitory activities for experimental verification, as well as predictions of their anticancer, pharmacokinetic, and toxic properties; oral bioavailability; and endocrine disruption using SwissADMET, PASS, ProTox-II, and Endocrine Disruptome web servers. A total of 12 metabolites, with a majority of flavonoids, were identified. Rutin, quercetin, and luteolin flavonoids demonstrated the highest ranked docking scores of -12.39, -11.15, and -10.43, respectively, while the half-maximal inhibitory concentration (IC50) values of these metabolites against XO activity were 11.35 µM, 11.1 µM, and 21.58 µM, respectively. In addition, SwissADMET generated data related to the physicochemical properties and drug-likeness of the metabolites. Similarly, the PASS, ProTox-II, and Endocrine Disruptome prediction models stated the safe and potential use of these natural compounds. However, in vivo studies are necessary to support the development of the prominent and promising therapeutic use of A. hierochuntica methanolic-leaf-extract-derived metabolites as XO inhibitors for the prevention and treatment of hyperuricemic and gout patients. Furthermore, the predicted findings of the present study open a new paradigm for these extract-derived metabolites by revealing novel oncogenic targets for the potential treatment of human malignancies.

PMID:39057691 | DOI:10.3390/metabo14070368

Metabolites. 2024 Jun 28;14(7):365. doi: 10.3390/metabo14070365.

ABSTRACT

Stoichiometric genome-scale metabolic models (generally abbreviated GSM, GSMM, or GEM) have had many applications in exploring phenotypes and guiding metabolic engineering interventions. Nevertheless, these models and predictions thereof can become limited as they do not directly account for protein cost, enzyme kinetics, and cell surface or volume proteome limitations. Lack of such mechanistic detail could lead to overly optimistic predictions and engineered strains. Initial efforts to correct these deficiencies were by the application of precursor tools for GSMs, such as flux balance analysis with molecular crowding. In the past decade, several frameworks have been introduced to incorporate proteome-related limitations using a genome-scale stoichiometric model as the reconstruction basis, which herein are called resource allocation models (RAMs). This review provides a broad overview of representative or commonly used existing RAM frameworks. This review discusses increasingly complex models, beginning with stoichiometric models to precursor to RAM frameworks to existing RAM frameworks. RAM frameworks are broadly divided into two categories: coarse-grained and fine-grained, with different strengths and challenges. Discussion includes pinpointing their utility, data needs, highlighting framework strengths and limitations, and appropriateness to various research endeavors, largely through contrasting their mathematical frameworks. Finally, promising future applications of RAMs are discussed.

PMID:39057688 | DOI:10.3390/metabo14070365

Curr Issues Mol Biol. 2024 Jul 4;46(7):6960. doi: 10.3390/cimb46070415.

ABSTRACT

Afiq Adham Abd Rasib and Roohaida Othman were not included as authors in the original publication [...].

PMID:39057102 | DOI:10.3390/cimb46070415

Cells. 2024 Jul 9;13(14):1171. doi: 10.3390/cells13141171.

ABSTRACT

Androgen excess is a key feature of several clinical phenotypes of polycystic ovary syndrome (PCOS). However, the presence of FSH receptor (FSHR) and aromatase (CYP19A1) activity responses to physiological endocrine stimuli play a critical role in the pathogenesis of PCOS. Preliminary data suggest that myo-Inositol (myo-Ins) and D-Chiro-Inositol (D-Chiro-Ins) may reactivate CYP19A1 activity. We investigated the steroidogenic pathway of Theca (TCs) and Granulosa cells (GCs) in an experimental model of murine PCOS induced in CD1 mice exposed for 10 weeks to a continuous light regimen. The effect of treatment with different combinations of myo-Ins and D-Chiro-Ins on the expression of Fshr, androgenic, and estrogenic enzymes was analyzed by real-time PCR in isolated TCs and GCs and in ovaries isolated from healthy and PCOS mice. Myo-Ins and D-Chiro-Ins, at a ratio of 40:1 at pharmacological and physiological concentrations, positively modulate the steroidogenic activity of TCs and the expression of Cyp19a1 and Fshr in GCs. Moreover, in vivo, inositols (40:1 ratio) significantly increase Cyp19a1 and Fshr. These changes in gene expression are mirrored by modifications in hormone levels in the serum of treated animals. Myo-Ins and D-Chiro-Ins in the 40:1 formula efficiently rescued PCOS features by up-regulating aromatase and FSHR levels while down-regulating androgen excesses produced by TCs.

PMID:39056753 | DOI:10.3390/cells13141171

Biology (Basel). 2024 Jul 9;13(7):512. doi: 10.3390/biology13070512.

ABSTRACT

Single-cell transcriptomics (scRNA-seq) is revolutionizing biological research, yet it faces challenges such as inefficient transcript capture and noise. To address these challenges, methods like neighbor averaging or graph diffusion are used. These methods often rely on k-nearest neighbor graphs from low-dimensional manifolds. However, scRNA-seq data suffer from the 'curse of dimensionality', leading to the over-smoothing of data when using imputation methods. To overcome this, sc-PHENIX employs a PCA-UMAP diffusion method, which enhances the preservation of data structures and allows for a refined use of PCA dimensions and diffusion parameters (e.g., k-nearest neighbors, exponentiation of the Markov matrix) to minimize noise introduction. This approach enables a more accurate construction of the exponentiated Markov matrix (cell neighborhood graph), surpassing methods like MAGIC. sc-PHENIX significantly mitigates over-smoothing, as validated through various scRNA-seq datasets, demonstrating improved cell phenotype representation. Applied to a multicellular tumor spheroid dataset, sc-PHENIX identified known extreme phenotype states, showcasing its effectiveness. sc-PHENIX is open-source and available for use and modification.

PMID:39056705 | DOI:10.3390/biology13070512

Biology (Basel). 2024 Jun 28;13(7):481. doi: 10.3390/biology13070481.

ABSTRACT

Triple-negative breast cancer (TNBC) cells are often resistant to FAS (CD95)-mediated apoptosis, but the underlying molecular mechanism(s) is not fully understood yet. Notably, the expression of the type II transmembrane protein, CD74, is correlated with chemotherapy-resistant and more invasive forms of cancers via unknown mechanisms. Here, we analyzed gene expression pattern of cancer patients and/or patient-derived xenograft (PDX) models and found that mRNA and protein levels of CD74 are highly expressed in TNBC and correlated with cancer stem cells (CSCs) and epithelial-mesenchymal transition (EMT) properties. Mechanistically, we found that AKT activation is likely critical for maintaining CD74 expression and protein stability to favor its oncogenic functions. Physiologically, epidermal growth factor (EGF) along with CD74 could activate AKT signaling, likely through binding of phosphorylated AKT (S473) to CD74, whereas inhibition of AKT could impair stability of CD74. We also revealed that CD74 binds to FAS and interferes with the intrinsic signaling of FAS-mediated apoptosis. As such, selective targeting of the CD74/FAS complex using the AKT inhibitor along with the CD74-derived peptide could synergistically restore and activate FAS-mediated apoptosis. Therefore, our approach of mobilizing apoptosis pathways likely provides a rationale for TNBC treatment by targeting the CD74/FAS and CD74-AKT axes.

PMID:39056676 | DOI:10.3390/biology13070481

Plant Cell. 2024 Jul 26:koae202. doi: 10.1093/plcell/koae202. Online ahead of print.

ABSTRACT

The domestication of crops, coupled with agroecosystem development, is associated with major environmental changes and provides an ideal model of phenotypic plasticity. Here, we examined 32 genotypes of three tetraploid wheat (Triticum turgidum L.) subspecies, wild emmer, emmer and durum wheat, which are representative of the key stages in the domestication of tetraploid wheat. We developed a pipeline that integrates RNA-Seq data and population genomics to assess gene expression plasticity and identify selection signatures under diverse nitrogen availability conditions. Our analysis revealed differing gene expression responses to nitrogen availability across primary (wild emmer to emmer) and secondary (emmer to durum wheat) domestication. Notably, nitrogen triggered the expression of twice as many genes in durum wheat compared to that in emmer and wild emmer. Unique selection signatures were identified at each stage: primary domestication mainly influenced genes related to biotic interactions, whereas secondary domestication affected genes related to amino acid metabolism, in particular lysine. Selection signatures were found in differentially expressed genes, notably those associated with nitrogen metabolism, such as the gene encoding glutamate dehydrogenase. Overall, our study highlights the pivotal role of nitrogen availability in the domestication and adaptive responses of a major food crop, with varying effects across different traits and growth conditions.

PMID:39056474 | DOI:10.1093/plcell/koae202

Front Dement. 2024;3:1332928. doi: 10.3389/frdem.2024.1332928. Epub 2024 Feb 12.

ABSTRACT

Alzheimer's disease (AD) is affecting a growing number of individuals. As a result, there is a pressing need for accurate and early diagnosis methods. This study aims to achieve this goal by developing an optimal data analysis strategy to enhance computational diagnosis. Although various modalities of AD diagnostic data are collected, past research on computational methods of AD diagnosis has mainly focused on using single-modal inputs. We hypothesize that integrating, or "fusing," various data modalities as inputs to prediction models could enhance diagnostic accuracy by offering a more comprehensive view of an individual's health profile. However, a potential challenge arises as this fusion of multiple modalities may result in significantly higher dimensional data. We hypothesize that employing suitable dimensionality reduction methods across heterogeneous modalities would not only help diagnosis models extract latent information but also enhance accuracy. Therefore, it is imperative to identify optimal strategies for both data fusion and dimensionality reduction. In this paper, we have conducted a comprehensive comparison of over 80 statistical machine learning methods, considering various classifiers, dimensionality reduction techniques, and data fusion strategies to assess our hypotheses. Specifically, we have explored three primary strategies: (1) Simple data fusion, which involves straightforward concatenation (fusion) of datasets before inputting them into a classifier; (2) Early data fusion, in which datasets are concatenated first, and then a dimensionality reduction technique is applied before feeding the resulting data into a classifier; and (3) Intermediate data fusion, in which dimensionality reduction methods are applied individually to each dataset before concatenating them to construct a classifier. For dimensionality reduction, we have explored several commonly-used techniques such as principal component analysis (PCA), autoencoder (AE), and LASSO. Additionally, we have implemented a new dimensionality-reduction method called the supervised encoder (SE), which involves slight modifications to standard deep neural networks. Our results show that SE substantially improves prediction accuracy compared to PCA, AE, and LASSO, especially in combination with intermediate fusion for multiclass diagnosis prediction.

PMID:39055313 | PMC:PMC11271260 | DOI:10.3389/frdem.2024.1332928