Biology subjects

Chu, Y.

Publications and source records attributed to Chu, Y..

At least 55 records · Page 3Linked to original sources

bioRxiv2022

Is Tau the Initial Pathology in Dopaminergic Nigrostriatal Degeneration? Studies in Parkinsonism and Parkinson ' Disease

While Parkinsons disease (PD) remains clinically defined by cardinal motor symptoms resulting from nigrostriatal degeneration, it is now appreciated that PD consists of multiple pathologies, but it is unclear which occurs first and which are responsible for the nigrostriatal degeneration. For the past number of years, we have been studying a well-characterized cohort of subjects with motor impairment that we have termed mild motor deficits (MMD). Motor deficits were determined on a modified and validated Unified Parkinsons Disease Rating Scale III (UPDRS III), but they occur to a degree insufficient to diagnose PD. We consider this population to have prodromal PD. However, in past studies, cases in this cohort had a selection bias as both a clinical syndrome in between no motor deficits and PD, plus nigral Lewy pathology as defined post-mortem, were required for inclusion. Therefore, in this study, we only based inclusion on a clinical phenotype intermediate between no motor impairment and PD. Then, we divided this group further based upon whether or not they had a synucleinopathy. Here we demonstrate that loss of nigral dopaminergic neurons, loss of putamenal dopaminergic innervation, loss of TH-phenotype in the substantia nigra and putamen, and changes in axonal transport occur equally in groups with and without nigral alpha-synuclein aggregates. Indeed, the common feature of these two groups is that both have similar degrees of AT8-expressing phospho-tau, a pathology not seen in the nigrostriatal system of aged-matched controls. These finding were confirmed with early (CP13) and late (PHF1) tau markers. This suggests that the initiation of nigrostriatal dopaminergic neurodegeneration occurs independently of alpha-synuclein aggregation and is likely tau mediated.

pathology↗

bioRxiv2022

A Self-attention Graph Convolutional Network for Precision Multi-tumour Early Diagnostics with DNA Methylation Data

DNA methylation data-based precision tumor early diagnostics is emerging as state of the art technology, which could capture the signals of cancer occurrence 3[~]5 years in advance and clinically more homogenous groups. At present, the sensitivity of early detection for many tumors is about 30%, which needs to be significantly improved. Nevertheless, based on the genome wide DNA methylation information, one could comprehensively characterize the entire molecular genetic landscape of the tumors and subtle differences among various tumors. With the accumulation of DNA methylation data, we need to develop high-performance methods that can model and consider more unbiased information. According to the above analysis, we have designed a self-attention graph convolutional network to automatically learn key methylation sites in a data-driven way for precision multi-tumor early diagnostics. Based on the selected methylation sites, we further trained a multi-class classification support vector machine. Large amount experiments have been conducted to investigate the performance of the computational pipeline. Experimental results demonstrated the effectiveness of the selected key methylation sites which are highly relevant for blood diagnosis.

bioinformatics↗

bioRxiv2022

Metabolic rescue ameliorates mitochondrial encephalo-cardiomyopathy in murine and human iPSC models of Leigh syndrome

Mice with deletion of complex I subunit Ndufs4 develop mitochondrial encephalomyopathy resembling Leigh syndrome (LS). We report that LS mice also develop severe cardiac bradyarrhythmia and diastolic dysfunction. Human induced pluripotent stem cell-derived cardiomyocytes (iPS-CMs) with Ndufs4 deletion recapitulate LS cardiomyopathy. Mechanistically, we demonstrate a direct link between complex I deficiency, decreased intracellular NAD+/ NADH and bradyarrhythmia, mediated by hyperacetylation of the cardiac sodium channel NaV1.5, particularly at K1479 site. Neuronal apoptosis in the cerebellar and midbrain regions in LS mice was associated with hyperacetylation of p53 and activation of microglia. Targeted metabolomics revealed increases in several amino acids and citric acid cycle intermediates, likely due to impairment of NAD+-dependent dehydrogenases, and a substantial decrease in reduced Glutathione (GSH). Metabolic rescue by nicotinamide riboside (NR) supplementation increased intracellular NAD+/ NADH, restored metabolic derangement, reversed protein hyperacetylation through NAD+-dependent Sirtuin deacetylase, and ameliorated cardiomyopathic phenotypes, concomitant with improvement of NaV1.5 current and SERCA2a function measured by Ca2+- transients. NR also attenuated neuronal apoptosis and microglial activation in the LS brain and human iPS-derived neurons with Ndufs4 deletion. Our study reveals direct mechanistic explanations of the observed cardiac bradyarrhythmia, diastolic dysfunction and neuronal apoptosis in mouse and human iPSC models of LS.

pathology↗

bioRxiv2021

Distribution of phosphorylated alpha-synuclein in non-diseased brain implicates olfactory bulb mitral cells in synucleinopathy pathogenesis

Synucleinopathies are neurodegenerative diseases characterized by pathological inclusions called "Lewy pathology" (LP) that consist of aggregated alpha-synuclein predominantly phosphorylated at serine 129 (PSER129). Despite the importance for understanding disease, little is known about the endogenous function of PSER129 or why it accumulates in disease. Here we conducted several observational studies using a sensitive tyramide signal amplification (TSA) technique to determine PSER129 distribution and function in the non-diseased mammalian brain. In wild-type non-diseased mice, PSER129 was detected in the olfactory bulb (OB) and several brain regions across the neuroaxis (i.e., OB to brain stem). In contrast, PSER129 immunoreactivity was not observed in any brain region of alpha-synuclein knockout mice. We found evidence of PSER129 positive structures in OB mitral cells of non-diseased mice, rats, non-human primates, and healthy humans. Using TSA multiplex fluorescent labeling we show that PSER129 positive punctate structures occur within inactive (i.e., cfos negative) T-box transcription factor 21 (TBX21) positive mitral cells and PSER129 within these cells was spatially associated with PK-resistant alpha-synuclein. Ubiquitin was found in PSER129 mitral cells but was not closely associated with PSER129. Biotinylation by antigen recognition (BAR) identified 125 PSER129-interacting proteins in the OB of healthy mice, which were significantly enriched for presynaptic vesicle trafficking/recycling, SNARE, fatty acid oxidation, oxidative phosphorylation, and RNA binding. TSA multiplex labeling confirmed the physical association of BAR identified protein Ywhag with PSER129 in the OB and in other regions across the neuroaxis. We conclude that PSER129 accumulates in mitral cells of the healthy OB as part of alpha-synuclein normal cellular functions. Incidental LP has been reported in the OB, and therefore we speculate that for synucleinopathies either; the disease processes begin locally in OB mitral cells or a systemic disease process is most apparent in the OB because the natural tendency to accumulate PSER129. Significance StatementMultiple lines of evidence have suggested that the disease process in some synucleinopathies begins in the olfactory bulb. Here we demonstrated that disease-associated phosphorylated alpha-synuclein preferentially occurs in mitral cells of the healthy mammalian olfactory bulb. We identified the protein interactome of phosphorylated alpha-synuclein in the healthy mouse olfactory bulb and established phosphorylated alpha-synuclein associates with presynaptic glutamatergic vesicles, SNARE machinery, and RNA metabolism machinery. Our data implicates olfactory bulb mitral cells in synucleinopathy pathogenesis. These findings advance our understanding of synucleinopathy disease origins and set the stage for new experimental models to interrogate the pathogenesis of synucleinopathies.

neuroscience↗

bioRxiv2021

Metabolic Derangement in Polycystic Kidney Disease Mouse Models Is Ameliorated by Mitochondrial-Targeted Antioxidants

Autosomal dominant polycystic kidney disease (ADPKD) is characterized by progressively enlarging cysts. Here we elucidate the interplay between oxidative stress, mitochondrial dysfunction, and metabolic derangement using two mouse models of PKD1 mutation, PKD1RC/null and PKD1RC/RC. Mouse kidneys with PKD1 mutation have decreased mitochondrial complexes activity. Targeted proteomics analysis shows a significant decrease in proteins involved in the TCA cycle, fatty acid oxidation (FAO), respiratory complexes, and endogenous antioxidants. Overexpressing mitochondrial-targeted catalase (mCAT) using adeno-associated virus reduces mitochondrial ROS, oxidative damage, ameliorates the progression of PKD and partially restores expression of proteins involved in FAO and the TCA cycle. In human ADPKD cells, inducing mitochondrial ROS increased ERK1/2 phosphorylation and decreased AMPK phosphorylation, whereas the converse was observed with increased scavenging of ROS in the mitochondria. Treatment with the mitochondrial protective peptide, SS31, recapitulates the beneficial effects of mCAT, supporting its potential application as a novel therapeutic for ADPKD.

pathology↗

bioRxiv2021

Sift-PULs: A public repository for specific functional polysaccharide utilization loci

BackgroundPolysaccharide utilization loci (PULs) were bacterial gene clusters encoding genes responsible for polysaccharide utilization process. PUL studies are blooming in recent years but the biochemical characterization speed is relative slow. There is a growing demand for PUL database with function annotations. ResultsUsing signature genes corresponding for specific polysaccharide, 10422 PULs specific for 6 polysaccharides (agar, alginate, pectin, carrageenan, chitin and {beta}-manan) from various bacterial phyla were predicted. Then online website of specific functional polysaccharide utilization loci (Sift-PULs) was constructed. Sift-PULs provides a repository where users could browse, search and download interested PULs without registration. ConclusionsThe key advantage of Sift-PULs is to assign a function annotation of each PUL, which is not available in existing PUL databases. PULs functional annotation lays a foundation for studying novel enzymes, new pathways, PUL evolution or bioengineering. The website is available on https://sift-puls.org

bioinformatics↗

bioRxiv2021

A highly conserved lipase deacylates oxidized phospholipids and ameliorates acute lung injury

Oxidized phospholipids have diverse biological activities, many of which can be pathological, yet how they are inactivated in vivo is not fully understood. Here we present evidence that a previously unsuspected host lipase, acyloxyacyl hydrolase (AOAH), can play a significant role in reducing the pro-inflammatory activities of two prominent products of phospholipid oxidation, 1-palmitoyl-2-glutaryl-sn-glycero-3- phosphocholine (PGPC) and 1-palmitoyl-2-(5-oxovaleroyl)-sn-glycero-3-phosphocholine (POVPC). AOAH removed the sn-2 and sn-1 acyl chains from both lipids and reduced their ability to induce macrophage inflammasome activation and cell death in vitro and acute lung injury in vivo. In addition to transforming Gram-negative bacterial lipopolysaccharide from stimulus to inhibitor, its most studied activity, AOAH can inactivate these important danger-associated molecular pattern (DAMP) molecules and reduce tissue inflammation and cell death.

immunology↗

bioRxiv2021

Dynamics of glioma-associated microglia and macrophages reveals their divergent roles in the immune response of brain

Glioma microenvironment contains numerous myeloid cells, including brain-resident microglia and recruited monocytes and macrophages (Mo/M{varphi}). When studied collectively, these cells presented pro-tumor effects. Yet, little is known about the differences among these myeloid populations. Using single-cell sequencing analysis, we studied the phenotypic characteristics, spatial variances, and dynamic changes of these relatively heterogeneous cell populations. Microglia populations with distinct spatial distribution presented different functional states, including tumor-associated subsets with phagocytic and lipid metabolism signature. Notably, this subset of glioma-associated microglia shared similar trait in a diverse spectrum of neuropathogenesis. In contrast, Mo/M{varphi} highly expressed genes related to angiogenesis, tumor invasion, and immune evasion. Moreover, identifying the Mo/M{varphi} subsets had prognostic and classificatory value in clinical application. These results thus eliminate the long-existing ambiguity about the role of microglia and Mo/M{varphi} in glioma pathogenesis, and reveal their prognostic and therapeutic value for glioma patients.

immunology↗

bioRxiv2021

Identification of a Major Locus for Flowering Pattern Sheds Light on Plant Architecture Diversification in Cultivated Peanut

Flowering pattern is a major taxonomic characteristic differentiating the two main subspecies of cultivated peanut (Arachis hypogaea L.). subsp. fastigiata possessing flowers on the mainstem (MSF) and a sequential flowering pattern, whereas subsp. hypogaea lacks flowers on the mainstem and exhibits an alternate flowering pattern. This character is considered the main contributor to plant architecture and the adaptability of each subgroup to specific growing conditions. Evidence indicates that flowering pattern differentiation occurred during the several thousand years of domestication and diversification in South America. However, the exact genetic mechanism that controls flowering pattern and the molecular changes that led to its historical diversification in peanut are unknown. We investigated the genetics of the flowering pattern in a recombinant inbred population of 259 lines (RILs), derivatives of an A. hypogaea and A. fastigiata cross. RILs segregated 1:1 in both the sequential/alternative and the MSF-plus/MSF-minus traits, indicating a single gene effect. Using the Axiom_Arachis2 SNP-array, MSF was mapped to a 1.7 Mbp segment on chromosome B02 of the cultivated A. hypogaea. Significant haplotype conservation was found for this locus in the USA peanut mini core collection, suggesting a possible selection upon hypogaea/fastigiata speciation. Furthermore, a candidate Terminal Flowering 1-like (AhTFL1) gene was identified within the MSF region, in which a 1492 bp deletion occurred in the fastigiata line that leads to a truncated protein product. Remapping MSF in the RIL population with the AhTFL1 deletion as a marker increased the LOD score from 53.3 to 158.8 with no recombination. The same deletion was also found to co-segregate with the phenotype in two EMS-mutagenized M2 families, suggesting a hotspot for large mutational deletion or gene conversion that may play a role in evolution. BLASTX analysis showed that the most similar homologous gene for TFL1-like in soybean is Det1, which previously was shown to control shoot determination. Sequence analysis of the TFL-1 in a series of domesticated lines showed that TFL1 was subjected to gain/loss events of the deletion, partly explaining the evolution of MSF in Arachis. Altogether, these results support the role of AhTFL-1 in peanut speciation during domestication and modern cultivation.

genetics↗

bioRxiv2021

The brown algae Saccharina japonica and Sargassum horneri exhibit species-specific responses to synergistic stress of ocean acidification and eutrophication

Ocean acidification and eutrophication are two important environmental stressors. They inevitably impact marine macroalgae, and hence the coastal ecosystem of China. Saccharina japonica, as the main culture species in China, is suffering the harmful golden tide caused by Sargassum horneri. However, it remains unclear whether the detrimental effects of S. horneri on S. japonica cultivation become more severe in future acidified and eutrophic scenario. In this study, we respectively investigated the effects of pCO2 (400 atm and 1000 atm) and nutrients (non-enriched and enriched seawater) on the growth, photosynthesis, respiration, chlorophyll contents, and tissue nitrogen of S. japonica and S. horneri. Results indicated that enrichment of nutrients contributed S. horneri to utilize HCO3-. The carbon acquisition pathway shifted from HCO3- to CO2 in S. japonica, while S. horneri remained using HCO3- regulated by nutrient enrichment. S. horneri exhibited better photosynthetic traits than S. japonica, with a higher level of net photosynthetic rate and chlorophyll contents at elevated pCO2 and enriched nutrients. Tissue nitrogen also accumulated richly in the thalli of S. horneri under higher pCO2 and nutrients. Significant enhancement in growth was only detected in S. horneri under synergistic stress. Together, S. horneri showed competitive dominance in current study. These findings suggest that increasing risk of golden tide in acidified and eutrophic ocean can most likely result in great damage to S. japonica cultivation.

ecology↗

bioRxiv2021

Accurate analysis of short read sequencing in complex genomes: A case study using QTL-seq to target blanchability in peanut (Arachis hypogaea)

Next Generation sequencing was a step change for molecular genetics and genomics. Illumina sequencing in particular still provides substantial value to animal and plant genomics. A simple yet powerful technique, referred to as QTL sequencing (QTL-seq) is susceptible to high levels of noise due to ambiguity of alignment of short reads in complex regions of the genome. This noise is particularly high when working with polyploid and/or outcrossing crop species, which impairs the efficacy of QTL-seq in identifying functional variation. By filtering loci based on the optimal alignment of short reads, we have developed a pipeline, named Khufu, that substantially improves the accuracy of QTL-seq analysis in complex genomes, allowing de novo variant discovery directly from bulk sequence. We first demonstrate the pipeline by identifying and validating loci contributing to blanching percentage in peanut using lines from multiple related populations. Using other published datasets in peanut, Brassica rapa, Hordeum volgare, Lactua satvia, and Felis catus, we demonstrate that Khufu produces more accurate results straight from bulk sequence. Khufu works across species, genome ploidy level, and data types. In cases where identified QTL were fine mapped, the fine mapped region corresponds to the top of the peak identified by Khufu. The accuracy of Khufu allows the analysis of population sequencing at very low coverage (<3x), greatly decreasing the amount of sequence needed to genotype even the most complex genomes.

genomics↗

bioRxiv2021

Impact of scaffolding protein TNRC6 paralogs on gene expression and splicing

TNRC6 is a scaffolding protein that bridges interactions between small RNAs, argonaute (AGO) protein, and effector proteins to control gene expression. There are three paralogs in mammalian cells, TNRC6A, TNRC6B, and TNRC6C. These paralogs have ~40% amino acid sequence identity and the extent of their unique or redundant functions is unclear. Here, we use knockout cell lines, enhanced crosslinking immunoprecipitation (eCLIP), and high-throughput RNA sequencing (RNAseq) to explore the roles of TNRC6 paralogs in RNA-mediated control of gene expression. We find that that the paralogs are largely functionally redundant and changes in levels of gene expression are well-correlated with those observed in AGO knockout cell lines. Splicing changes observed in AGO knockout cell lines are observed in TNRC6 knockout cells. These data further define the roles of the TNRC6 isoforms as part of the RNA interference (RNAi) machinery.

molecular biology↗

bioRxiv2021

Argonaute Binding within Human Nuclear RNA and its Impact on Alternative Splicing

Mammalian RNA interference (RNAi) is often linked to the regulation of gene expression in the cytoplasm. Synthetic RNAs, however, can also act through the RNAi pathway to regulate transcription and splicing. While nuclear regulation by synthetic RNAs can be robust, a critical unanswered question is whether endogenous functions for nuclear RNAi exist in mammalian cells. Using enhanced crosslinking immunoprecipitation (eCLIP) in combination with RNA sequencing (RNAseq) and multiple AGO knockout cell lines, we mapped AGO2 protein binding sites within nuclear RNA. The strongest AGO2 binding sites were mapped to micro RNAs (miRNAs). The most abundant miRNAs were distributed similarly between cytoplasm and nucleus, providing no evidence for mechanisms that facilitate localization of miRNAs in one compartment versus the other. Beyond miRNAs, most statistically-significant AGO2 binding was within introns. Splicing changes were confirmed by RT-PCR and were recapitulated by synthetic duplex RNAs and miRNA mimics complementary to the sites of AGO2 binding. These data support the hypothesis that miRNAs can control gene splicing. While nuclear RNAi proteins have the potential to be a natural regulatory mechanism, careful study will be necessary to identify critical RNA drivers of normal physiology and disease.

molecular biology↗

bioRxiv2020

Detection And Purification Of Lewy Pathology From Formalin Fixed Primary Human Tissue Using Biotinylation By Antigen Recognition

The intracellular misfolding and accumulation of alpha-synuclein into structures collectively called Lewy pathology is a central phenomenon for the pathogenesis of Parkinsons disease (PD), Dementia with Lewy Bodies (DLB), and Multiple System Atrophy. Understanding the molecular architecture of Lewy pathology is crucial for understanding disease origins and progression. Here we developed a method to label, extract, and purify molecules from Lewy pathology of formalin fixed PD and DLB brain for blotting and mass spectrometry analysis. Using the biotinylation antibody recognition (BAR) technique, we labeled phosphoserine 129 alpha-synuclein positive pathology and associated molecules with biotin. Formalin crosslinks were then reversed, protein extracted, and pathology associated molecules isolated with streptavidin beads. Results showed superior immunohistochemical staining of Lewy pathology following the BAR protocol when compared to standard avidin biotin complex (ABC) based detection. The enhanced staining was particularly apparent for fibers of the medial forebrain bundle and punctate pathology within the striatum and cortex, which otherwise were weakly labeled or not detected. Subsequent immunoblotting BAR-labeled Lewy pathology extracts revealed the presence of high molecular weight alpha-synuclein, ubiquitin protein conjugates, and phosphoserine 129 alpha-synuclein. Mass spectrometry analysis of BAR-labeled Lewy pathology extracts from PD and DLB patients identified 815 proteins with significant enrichment for many pathways. Notably the most significant KEGG pathway was Parkinsons disease (FDR = 2.48 x 10-26) and GO Cellular compartment was extracellular exosomes (GO Cellular Compartment; FDR = 2.66x 10-34). We used enrichment data to create a functional map of Lewy Pathology from primary disease tissues, which implicated Vesicle Trafficking as the primary disease associated pathway in DLB and PD. In summary, this protocol can be used to enrich for Lewy pathology from formalin fixed human primary tissues, which allows the determination of molecular signatures of Lewy pathology. This technique has broad potential to help understand the phenomenon of Lewy pathology in primary human tissue and animal models.

neuroscience↗

bioRxiv2020

Genetic determinants of COVID-19 drug efficacy revealed by genome-wide CRISPR screens

Immunomodulatory agents dexamethasone and colchicine, antiviral drugs remdesivir, favipiravir and ribavirin, as well as antimalarial drugs chloroquine phosphate and hydroxychloroquine are currently used in the combat against COVID-191-16. However, whether some of these drugs have clinical efficacy for COVID-19 is under debate. Moreover, these drugs are applied in COVID-19 patients with little knowledge of genetic biomarkers, which will hurt patient outcome. To answer these questions, we designed a screen approach that could employ genome-wide sgRNA libraries to systematically uncover genes crucial for these drugs action. Here we present our findings, including genes crucial for the import, export, metabolic activation and inactivation of remdesivir, as well as genes that regulate colchicine and dexamethasones immunosuppressive effects. Our findings provide preliminary information for developing urgently needed genetic biomarkers for these drugs. Such biomarkers will help better interpret COVID-19 clinical trial data and point to how to stratify COVID-19 patients for proper treatment with these drugs.

pharmacology and toxicology↗

bioRxiv2020

Multivariate genome-wide association study identifies 1735 unique genetic loci associated with cortical morphology.

Brain morphology has been shown to be highly heritable, yet only a small portion of the heritability is explained by the genetic variants discovered so far. Here we exploit the distributed nature of genetic effects across the brain and apply the Multivariate Omnibus Statistical Test (MOSTest) to genome-wide association studies (GWAS) of vertex-wise structural magnetic resonance imaging (MRI) cortical measures from N=35,657 participants in the UK Biobank. We identified 695 loci for cortical surface area and 539 for cortical thickness, in total 780 unique genetic loci associated with cortical morphology. This reflects an approximate 10-fold increase compared to the commonly applied univariate GWAS methods. Power analysis indicates that applying MOSTest to vertex-wise structural MRI data triples the effective sample size compared to conventional univariate GWAS approaches. Functional follow up including gene-based analyses implicate 10% of all protein-coding genes and point towards pathways involved in neurogenesis and cell differentiation.

genetics↗

bioRxiv2020

A neural network-based framework to understand the Type 2 Diabetes (T2D)-related alteration of the human gut microbiome

To identify the microbial markers from the complex human gut microbiome for delineating the disease-related microbial alteration is of great interest. Here, we develop a framework combining neural network (NN) and random forest (RF), resulting in 40 marker species and 90 marker genes identified from the metagenomic dataset D1 (185 healthy and 183 type 2 diabetes (T2D) samples), respectively. Using these markers, the NN model obtains higher accuracy in classifying the T2D-related samples than machine learning-based approaches. The NN-based regression analysis determines the fasting blood glucose (FBG) is the most significant association factor (P<<0.05) in the T2D-related alteration of the gut microbiome. Twenty-four marker species that vary little across the case and control samples and are often neglected by the statistic-based methods greatly shift in different stages of the T2D development, implying that the cumulative effect of the markers rather than individuals likely drives the alteration of the gut microbiome.

bioinformatics↗

bioRxiv2020

Predicting drug-target interactions using multi-label learning with community detection method (DTI-MLCD)

Identifying drug-target interactions (DTIs) is an important step for drug discovery and drug repositioning. To reduce heavily experiment cost, booming machine learning has been applied to this field and developed many computational methods, especially binary classification methods. However, there is still much room for improvement in the performance of current methods. Multi-label learning can reduce difficulties faced by binary classification learning with high predictive performance, and has not been explored extensively. The key challenge it faces is the exponential-sized output space, and considering label correlations can help it. Thus, we facilitate the multi-label classification by introducing community detection methods for DTIs prediction, named DTI-MLCD. On the other hand, we updated the gold standard data set proposed in 2008 and still in use today. The proposed DTI-MLCD is performed on the gold standard data set before and after the update, and shows the superiority than other classical machine learning methods and other benchmark proposed methods, which confirms the efficiency of it. The data and code for this study can be found at https://github.com/a96123155/DTI-MLCD.

bioinformatics↗