Chest radiography deep radiomics-enabled aortic arch calcification interpretation across different populations.

Earlier detection of aortic calcification can facilitate subsequent cardiovascular care planning. Opportunistic screening based on plain chest radiography is potentially feasible in various population. We used multiple deep convolutional neural network (CNN) transfer learning by fine-tuning pre-trained models followed by ensemble technique for aortic arch calcification on chest radiographs from a derivation and two external databases with distinct features. Our ensemble approach achieved 84.12% precision, 84.70% recall, and an area under the receiver-operating-characteristic curve (AUC) of 0.85 in the general population/older adult's dataset. We also obtained 87.5% precision, 85.56% recall, and an AUC of 0.86 in the pre-end-stage kidney disease (pre-ESKD) cohort. We identified discriminative regions for identifying aortic arch calcification between patients without and with pre-ESKD. These findings are expected to optimize cardiovascular risk prediction if our model is incorporated into the process of routine care.

The context-dependent role of transforming growth factor-ß/miR-378a-3p/connective tissue growth factor in vascular calcification: a translational study.

BACKGROUND: Vascular calcification (VC) constitutes an important vascular pathology with prognostic importance. The pathogenic role of transforming growth factor-ß (TGF-ß) in VC remains unclear, with heterogeneous findings that we aimed to evaluate using experimental models and clinical specimens. METHODS: Two approaches, exogenous administration and endogenous expression upon osteogenic media (OM) exposure, were adopted. Aortic smooth muscle cells (ASMCs) were subjected to TGF-ß1 alone, OM alone, or both, with calcification severity determined. We evaluated miR-378a-3p and TGF-ß1 effectors (connective tissue growth factor; CTGF) at different periods of calcification. Results were validated in an ex vivo model and further in sera from older adults without or with severe aortic arch calcification. RESULTS: TGF-ß1 treatment induced a significant dose-responsive increase in ASMC calcification without or with OM at the mature but not early or mid-term VC period. On the other hand, OM alone induced VC accompanied by suppressed TGF-ß1 expressions over time; this phenomenon paralleled the declining miR-378a-3p and CTGF expressions since early VC. TGF-ß1 treatment led to an upregulation of CTGF since early VC but not miR-378a-3p until mid-term VC, while miR-378a-3p overexpression suppressed CTGF expressions without altering TGF-ß1 levels. The OM-induced down-regulation of TGF-ß1 and CTGF was also observed in the ex vivo models, with compatible results identified from human sera. CONCLUSIONS: We showed that TGF-ß1 played a context-dependent role in VC, involving a time-dependent self-regulatory loop of TGF-ß1/miR-378a-3p/CTGF signaling. Our findings may assist subsequent studies in devising potential therapeutics against VC.

Deep Learning-Assisted Repurposing of Plant Compounds for Treating Vascular Calcification: An In Silico Study with Experimental Validation.

BACKGROUND: Vascular calcification (VC) constitutes subclinical vascular burden and increases cardiovascular mortality. Effective therapeutics for VC remains to be procured. We aimed to use a deep learning-based strategy to screen and uncover plant compounds that potentially can be repurposed for managing VC. METHODS: We integrated drugome, interactome, and diseasome information from Comparative Toxicogenomic Database (CTD), DrugBank, PubChem, Gene Ontology (GO), and BioGrid to analyze drug-disease associations. A deep representation learning was done using a high-level description of the local network architecture and features of the entities, followed by learning the global embeddings of nodes derived from a heterogeneous network using the graph neural network architecture and a random forest classifier established for prediction. Predicted results were tested in an in vitro VC model for validity based on the probability scores. RESULTS: We collected 6,790 compounds with available Simplified Molecular-Input Line-Entry System (SMILES) data, 11,958 GO terms, 7,238 diseases, and 25,482 proteins, followed by local embedding vectors using an end-to-end transformer network and a node2vec algorithm and global embedding vectors learned from heterogeneous network via the graph neural network. Our algorithm conferred a good distinction between potential compounds, presenting as higher prediction scores for the compound categories with a higher potential but lower scores for other categories. Probability score-dependent selection revealed that antioxidants such as sulforaphane and daidzein were potentially effective compounds against VC, while catechin had low probability. All three compounds were validated in vitro. CONCLUSIONS: Our findings exemplify the utility of deep learning in identifying promising VC-treating plant compounds. Our model can be a quick and comprehensive computational screening tool to assist in the early drug discovery process.

DNA methylome and transcriptome landscapes of cancer-associated fibroblasts reveal a smoking-associated malignancy index.

Unlike the better-studied aberrant epigenome in the tumor, the clinicopathologic impact of DNA methylation in the tumor microenvironment (TME), especially the contribution from cancer-associated fibroblasts (CAFs), remains elusive. CAFs exhibit profound patient-to-patient tumorigenic heterogeneity. We asked whether such heterogeneity may be exploited to quantify the level of TME malignancy. We developed a robust and efficient methylome/transcriptome co-analytical system for CAFs and paired normal fibroblasts (NFs) from non-small-cell lung cancer patients. We found 14,781 CpG sites of CAF/NF differential methylation, of which 3,707 sites showed higher methylation changes in ever-smokers than in nonsmokers. Concomitant CAF/NF differential gene expression analysis pointed to a subset of 54 smoking-associated CpG sites with strong methylation-regulated gene expression. A methylation index that summarizes the ß values of these CpGs was built for NF/CAF discrimination (MIND) with high sensitivity and specificity. The potential of MIND in detecting premalignancy across individual patients was shown. MIND succeeded in predicting tumor recurrence in multiple lung cancer cohorts without reliance on patient survival data, suggesting that the malignancy level of TME may be effectively graded by this index. Precision TME grading may provide additional pathological information to guide cancer prognosis and open up more options in personalized medicine.

Superoxide Dismutase 2 (SOD2) in Vascular Calcification: A Focus on Vascular Smooth Muscle Cells, Calcification Pathogenesis, and Therapeutic Strategies.

Vascular calcification (VC) describes the pathophysiological phenotype of calcium apatite deposition within the vascular wall, leading to vascular stiffening and the loss of compliance. VC is never benign; the presence and severity of VC correlate closely with the risk of myocardial events and cardiovascular mortality in multiple at-risk populations such as patients with diabetes and chronic kidney disease. Mitochondrial dysfunction involving each of vascular wall constituents (endothelia and vascular smooth muscle cells (VSMCs)) aggravates various vascular pathologies, including atherosclerosis and VC. However, few studies address the pathogenic role of mitochondrial dysfunction during the course of VC, and mitochondrial reactive oxygen species (ROS) seem to lie in the pathophysiologic epicenter. Superoxide dismutase 2 (SOD2), through its preferential localization to the mitochondria, stands at the forefront against mitochondrial ROS in VSMCs and thus potentially modifies the probability of VC initiation or progression. In this review, we will provide a literature-based summary regarding the relationship between SOD2 and VC in the context of VSMCs. Apart from the conventional wisdom of attenuating mitochondrial ROS, SOD2 has been found to affect mitophagy and the formation of the autophagosome, suppress JAK/STAT as well as PI3K/Akt signaling, and retard vascular senescence, all of which underlie the beneficial influences on VC exerted by SOD2. More importantly, we outline the therapeutic potential of a novel SOD2-targeted strategy for the treatment of VC, including an ever-expanding list of pharmaceuticals and natural compounds. It is expected that VSMC SOD2 will become an important druggable target for treating VC in the future.

A combined microRNA and target protein-based panel for predicting the probability and severity of uraemic vascular calcification: a translational study.

AIMS: Vascular calcification (VC) increases the future risk of cardiovascular events in uraemic patients, but effective therapies are still unavailable. Accurate identification of those at risk of developing VC using pathogenesis-based biomarkers is of particular interest and may facilitate individualized risk stratification. We aimed to uncover microRNA (miRNA)-target protein-based biomarker panels for evaluating uraemic VC probability and severity. METHODS AND RESULTS: We created a three-tiered in vitro VC model and an in vivo uraemic rat model receiving high phosphate diet to mimic uraemic VC. RNAs from the three-tiered in vitro and in vivo uraemic VC models underwent miRNA and mRNA microarray, with results screened for differentially expressed miRNAs and their target genes as biomarkers. Findings were validated in original models and additionally in an ex vivo VC model and human cells, followed by functional assays of identified miRNAs and target proteins, and tests of sera from end-stage renal disease (ESRD) and non-dialysis-dependent chronic kidney disease (CKD) patients without and with VC. Totally 122 down-regulated and 119 up-regulated miRNAs during calcification progression were identified initially; further list narrowing based on miRNA-mRNA pairing, anti-correlation, and functional enrichment left 16 and 14 differentially expressed miRNAs and mRNAs. Levels of four miRNAs (miR-10b-5p, miR-195, miR-125b-2-3p, and miR-378a-3p) were shown to decrease throughout all models tested, while one mRNA (SULF1, a potential target of miR-378a-3p) exhibited the opposite trend concurrently. Among 96 ESRD (70.8% with VC) and 59 CKD patients (61% with VC), serum miR-125b2-3p and miR-378a-3p decreased with greater VC severity, while serum SULF1 levels increased. Adding serum miR-125b-2-3p, miR-378a-3p, and SULF1 into regression models for VC substantially improved performance compared to using clinical variables alone. CONCLUSION: Using a translational approach, we discovered a novel panel of biomarkers for gauging the probability/severity of uraemic VC based on miRNAs/target proteins, which improved the diagnostic accuracy.

The M1/M2 spectrum and plasticity of malignant pleural effusion-macrophage in advanced lung cancer.

BACKGROUND: Malignant pleural effusion (MPE)-macrophage (Mφ) of lung cancer patients within unique M1/M2 spectrum showed plasticity in M1-M2 transition. The M1/M2 features of MPE-Mφ and their significance to patient outcomes need to be clarified; furthermore, whether M1-repolarization could benefit treatment remains unclear. METHODS: Total 147 stage-IV lung adenocarcinoma patients undergoing MPE drainage were enrolled for profiling and validation of their M1/M2 spectrum. In addition, the MPE-Mφ signature on overall patient survival was analyzed. The impact of the M1-polarization strategy of patient-derived MPE-Mφ on anti-cancer activity was examined. RESULTS: We found that MPE-Mφ expressed both traditional M1 (HLA-DRA) and M2 (CD163) markers and showed a wide range of M1/M2 spectrum. Most of the MPE-Mφ displayed diverse PD-L1 expression patterns, while the low PD-L1 expression group was correlated with higher levels of IL-10. Among these markers, we identified a novel two-gene MPE-Mφ signature, IL-1ß and TGF-ß1, representing the M1/M2 tendency, which showed a strong predictive power in patient outcomes in our MPE-Mφ patient cohort (N = 60, p = 0.013) and The Cancer Genome Atlas Lung Adenocarcinoma dataset (N = 478, p < 0.0001). Significantly, ß-glucan worked synergistically with IFN-γ to reverse the risk signature by repolarizing the MPE-Mφ toward the M1 pattern, enhancing anti-cancer activity. CONCLUSIONS: We identified MPE-Mφ on the M1/M2 spectrum and plasticity and described a two-gene M1/M2 signature that could predict the outcome of late-stage lung cancer patients. In addition, we found that "re-education" of these MPE-Mφ toward anti-cancer M1 macrophages using clinically applicable strategies may overcome tumor immune escape and benefit anti-cancer therapies.

Astaxanthin Counteracts Vascular Calcification In Vitro Through an Early Up-Regulation of SOD2 Based on a Transcriptomic Approach.

Vascular calcification (VC) is a critical contributor to the rising cardiovascular risk among at-risk populations such as those with diabetes or renal failure. The pathogenesis of VC involves an uprising of oxidative stress, for which antioxidants can be theoretically effective. However, astaxanthin, a potent antioxidant, has not been tested before for the purpose of managing VC. To answer this question, we tested the efficacy of astaxanthin against VC using the high phosphate (HP)-induced vascular smooth muscle cell (VSMC) calcification model. RNAs from treated groups underwent Affymetrix microarray screening, with intra-group consistency and inter-group differential expressions identified. Candidate hub genes were selected, followed by validation in experimental models and functional characterization. We showed that HP induced progressive calcification among treated VSMCs, while astaxanthin dose-responsively and time-dependently ameliorated calcification severities. Transcriptomic profiling revealed that 3491 genes exhibited significant early changes during VC progression, among which 26 potential hub genes were selected based on closeness ranking and biologic plausibility. SOD2 was validated in the VSMC model, shown to drive the deactivation of cellular senescence and enhance antioxidative defenses. Astaxanthin did not alter intracellular reactive oxygen species (ROS) levels without HP, but significantly lowered ROS production in HP-treated VSMCs. SOD2 knockdown prominently abolished the anti-calcification effect of astaxanthin on HP-treated VSMCs, lending support to our findings. In conclusion, we demonstrated for the first time that astaxanthin could be a potential candidate treatment for VC, through inducing the up-regulation of SOD2 early during calcification progression and potentially suppressing vascular senescence.

Predicting the Associations between Meridians and Chinese Traditional Medicine Using a Cost-Sensitive Graph Convolutional Neural Network.

Natural products are the most important and commonly used in Traditional Chinese Medicine (TCM) for healthcare and disease prevention in East-Asia. Although the Meridian system of TCM was established several thousand years ago, the rationale of Meridian classification based on the ingredient compounds remains poorly understood. A core challenge for the traditional machine learning approaches for chemical activity prediction is to encode molecules into fixed length vectors but ignore the structural information of the chemical compound. Therefore, we apply a cost-sensitive graph convolutional neural network model to learn local and global topological features of chemical compounds, and discover the associations between TCM and their Meridians. In the experiments, we find that the performance of our approach with the area under the receiver operating characteristic curve (ROC-AUC) of 0.82 which is better than the traditional machine learning algorithm and also obtains 8%-13% improvement comparing with the state-of-the-art methods. We investigate the powerful ability of deep learning approach to learn the proper molecular descriptors for Meridian prediction and to provide novel insights into the complementary and alternative medicine of TCM.

Natural and non-natural antioxidative compounds: potential candidates for treatment of vascular calcification.

Vascular calcification (VC) is highly prevalent in patients with advanced age, or those with chronic kidney disease and diabetes, accounting for substantial global cardiovascular burden. The pathophysiology of VC involves active mineral deposition by transdifferentiated vascular smooth muscle cells exhibiting osteoblast-like behavior, building upon cores with or without apoptotic bodies. Oxidative stress drives the progression of the cellular phenotypic switch and calcium deposition in the vascular wall. In this review, we discuss potential compounds that shield these cells from the detrimental influences of reactive oxygen species as promising treatment options for VC. A comprehensive summary of the current literature regarding antioxidants for VC is important, as no effective therapy is currently available for this disease. We systematically searched through the existing literature to identify original articles investigating traditional antioxidants and novel compounds with antioxidant properties with regard to their effectiveness against VC in experimental or clinical settings. We uncovered 36 compounds with antioxidant properties against VC pathology, involving mechanisms such as suppression of NADPH oxidase, BMP-2, and Wnt/ß-catenin; anti-inflammation; and activation of Nrf2 pathways. Only two compounds have been tested clinically. These findings suggest that a considerable opportunity exists to harness these antioxidants for therapeutic use for VC. In order to achieve this goal, more translational studies are needed.

MicroRNA-125b in vascular diseases: An updated systematic review of pathogenetic implications and clinical applications.

Epigenetic changes, particularly non-coding RNAs, have been implicated extensively in the pathogenesis of vascular diseases. Specific miRNAs are involved in the differentiation, phenotypic switch, proliferation, apoptosis, cytokine production and matrix deposition of endothelial cells and/or vascular smooth muscle cells. MicroRNA-125b has been studied in depth for its role in carcinogenesis with a double-edged role; that is, it can act as an oncogene in some cancer types and as a tumour suppressor gene in others. However, cumulative evidence from the use of advanced miRNA profiling techniques and bioinformatics analysis suggests that miR-125b can be a potential mediator and useful marker of vascular diseases. Currently, the exact role of miR-125b in vascular diseases is not known. In this systematic review, we intend to provide an updated compilation of all the recent findings of miR-125b in vascular diseases, using a systematic approach of retrieving data from all available reports followed by data summarization. MiR-125b serves as a pathogenic player in multiple vascular pathologies involving endothelia and vascular smooth muscle cells and also serves as a diagnostic marker for vascular diseases. We further provide a computational biologic presentation of the complex network of miR-125b and its target genes within the scope of vascular diseases.

Risk Factors Associated With Altered Circulating Micro RNA -125b and Their Influences on Uremic Vascular Calcification Among Patients With End-Stage Renal Disease.

Background Micro RNA -125b (miR-125b) has been shown to regulate vascular calcification ( VC ), and serum miR-125b levels are a potential biomarker for estimating the risk of uremic VC status. However, it is unknown whether clinical features, including chronic kidney disease-mineral bone disorder molecules, affect serum miR-125b levels. Methods and Results Patients receiving chronic dialysis for ≥3 months were recruited from different institutes. Serum miR-125b and chronic kidney disease-mineral bone disorder effectors, including intact parathyroid hormone, 25- OH -D, fibroblast growth factor-23, osteoprotegerin, and fetuin-A, were quantified. We used multivariate regression analyses to identify factors associated with low serum miR-125b levels and an area under receiver operating characteristic curve curve to derive optimal cutoffs for factors exhibiting close associations. Further regression analyses evaluated the influence of miR-125b on VC risk. Among 223 patients receiving chronic dialysis (mean age, 67.3 years; mean years of dialysis, 5.2), 54 (24.2%) had high serum miR-125b levels. Osteoprotegerin ( P=0.013), fibroblast growth factor-23 ( P=0.006), and fetuin-A ( P=0.036) were linearly associated with serum miR-125b levels. High osteoprotegerin levels independently correlated with high serum miR-125 levels. Adding serum miR-125b levels and serum osteoprotegerin levels (≥400 pg/mL) into models estimating the risk of uremic VC increased the area under receiver operating characteristic curve values (for models without miR-125b/osteoprotegerin, with miR-125b, and both: 0.74, 0.79, and 0.81, respectively). Conclusions Serum osteoprotegerin levels ≥400 pg/mL and serum miR-125b levels synergistically increased the accuracy of estimating VC risk among patients receiving chronic dialysis. Taking miR-125b and osteoprotegerin levels into consideration when estimating VC risk may be recommended.

Circulating MicroRNA-125b Predicts the Presence and Progression of Uremic Vascular Calcification.

OBJECTIVE: Vascular calcification (VC) is a major cause of mortality in patients with end-stage renal diseases. Biomarkers to predict the progression of VC early are in urgent demand. APPROACH AND RESULTS: We identified circulating, cell-free microRNAs as potential biomarkers using in vitro VC models in which both rat and human aortic vascular smooth muscle cells were treated with high levels of phosphate to mimic uremic hyperphosphatemia. Using an Affymetrix microRNA array, we found that miR-125b and miR-382 expression levels declined significantly as biomineralization progressed, but this decline was only observed for miR-125b in the culture medium. A time-dependent decrease in aortic tissue and serum miR-125b levels was also found in both ex vivo and in vivo renal failure models. We examined the levels of circulating, cell-free miR-125b in sera from patients with end-stage renal diseases (n=88) and found an inverse association between the severity of VC and the circulating miR-125b level, irrespective of age or mineral-related hormones (odds ratio, 0.71; P=0.03). Furthermore, serum miR-125b levels on enrollment can predict VC progression years later (for high versus low, odds ratio, 0.14; P<0.01; for the highest versus lowest tertile and middle versus lowest tertile, odds ratio, 0.55 and 0.13; P=0.3 and <0.01, respectively). The uremic VC prediction efficacy using circulating miR-125b levels was also observed in an independent cohort (n=135). CONCLUSIONS: The results suggest that serum miR-125b levels are associated with VC severity and serve as a novel predictive marker for the risk of uremia-associated calcification progression.

A brief survey on computational approaches to reveal drug and disease associations.

People worldwide are still threatened by various complex disease phenotypes, especially cancer which is usually caused by the accumulation of multi-factor-driven alterations. Although drugs achieve the therapeutic functions by targeting particular molecular, the therapies used nowadays against diseases are not effective enough due to the limitation of the knowledge about the drug-disease associations. The rapid increasing of the available experimental data and knowledge enable scientists to reveal drug-disease associations by the systematic integration and analysis. In this review, we show that several computational methods can help us to explain the underlying relationships between pharmacology and pathology. It is expected that newer computational methods will take advantage of heterogeneous and multi-dimensional data and increase the efficacy and safety of existing drugs for disease treatment.

Inferring drug-disease associations from integration of chemical, genomic and phenotype data using network propagation.

BACKGROUND: During the last few years, the knowledge of drug, disease phenotype and protein has been rapidly accumulated and more and more scientists have been drawn the attention to inferring drug-disease associations by computational method. Development of an integrated approach for systematic discovering drug-disease associations by those informational data is an important issue. METHODS: We combine three different networks of drug, genomic and disease phenotype and assign the weights to the edges from available experimental data and knowledge. Given a specific disease, we use our network propagation approach to infer the drug-disease associations. RESULTS: We apply prostate cancer and colorectal cancer as our test data. We use the manually curated drug-disease associations from comparative toxicogenomics database to be our benchmark. The ranked results show that our proposed method obtains higher specificity and sensitivity and clearly outperforms previous methods. Our result also show that our method with off-targets information gets higher performance than that with only primary drug targets in both test data. CONCLUSIONS: We clearly demonstrate the feasibility and benefits of using network-based analyses of chemical, genomic and phenotype data to reveal drug-disease associations. The potential associations inferred by our method provide new perspectives for toxicogenomics and drug reposition evaluation.

Biomarker identification for prostate cancer and lymph node metastasis from microarray data and protein interaction network using gene prioritization method.

Finding a genetic disease-related gene is not a trivial task. Therefore, computational methods are needed to present clues to the biomedical community to explore genes that are more likely to be related to a specific disease as biomarker. We present biomarker identification problem using gene prioritization method called gene prioritization from microarray data based on shortest paths, extended with structural and biological properties and edge flux using voting scheme (GP-MIDAS-VXEF). The method is based on finding relevant interactions on protein interaction networks, then scoring the genes using shortest paths and topological analysis, integrating the results using a voting scheme and a biological boosting. We applied two experiments, one is prostate primary and normal samples and the other is prostate primary tumor with and without lymph nodes metastasis. We used 137 truly prostate cancer genes as benchmark. In the first experiment, GP-MIDAS-VXEF outperforms all the other state-of-the-art methods in the benchmark by retrieving the truest related genes from the candidate set in the top 50 scores found. We applied the same technique to infer the significant biomarkers in prostate cancer with lymph nodes metastasis which is not established well.

Pathway detection from protein interaction networks and gene expression data using color-coding methods and A∗ search algorithms.

With the large availability of protein interaction networks and microarray data supported, to identify the linear paths that have biological significance in search of a potential pathway is a challenge issue. We proposed a color-coding method based on the characteristics of biological network topology and applied heuristic search to speed up color-coding method. In the experiments, we tested our methods by applying to two datasets: yeast and human prostate cancer networks and gene expression data set. The comparisons of our method with other existing methods on known yeast MAPK pathways in terms of precision and recall show that we can find maximum number of the proteins and perform comparably well. On the other hand, our method is more efficient than previous ones and detects the paths of length 10 within 40 seconds using CPU Intel 1.73 GHz and 1 GB main memory running under windows operating system.

Identifying co-targets to fight drug resistance based on a random walk model.

BACKGROUND: Drug resistance has now posed more severe and emergent threats to human health and infectious disease treatment. However, wet-lab approaches alone to counter drug resistance have so far still achieved limited success due to less knowledge about the underlying mechanisms of drug resistance. Our approach apply a heuristic search algorithm in order to extract active network under drug treatment and use a random walk model to identify potential co-targets for effective antibacterial drugs. RESULTS: We use interactome network of Mycobacterium tuberculosis and gene expression data which are treated with two kinds of antibiotic, Isoniazid and Ethionamide as our test data. Our analysis shows that the active drug-treated networks are associated with the trigger of fatty acid metabolism and synthesis and nicotinamide adenine dinucleotide (NADH)-related processes and those results are consistent with the recent experimental findings. Efflux pumps processes appear to be the major mechanisms of resistance but SOS response is significantly up-regulation under Isoniazid treatment. We also successfully identify the potential co-targets with literature confirmed evidences which are related to the glycine-rich membrane, adenosine triphosphate energy and cell wall processes. CONCLUSIONS: With gene expression and interactome data supported, our study points out possible pathways leading to the emergence of drug resistance under drug treatment. We develop a computational workflow for giving new insights to bacterial drug resistance which can be gained by a systematic and global analysis of the bacterial regulation network. Our study also discovers the potential co-targets with good properties in biological and graph theory aspects to overcome the problem of drug resistance.

A network flow approach to predict drug targets from microarray data, disease genes and interactome network - case study on prostate cancer.

BACKGROUND: Systematic approach for drug discovery is an emerging discipline in systems biology research area. It aims at integrating interaction data and experimental data to elucidate diseases and also raises new issues in drug discovery for cancer treatment. However, drug target discovery is still at a trial-and-error experimental stage and it is a challenging task to develop a prediction model that can systematically detect possible drug targets to deal with complex diseases. METHODS: We integrate gene expression, disease genes and interaction networks to identify the effective drug targets which have a strong influence on disease genes using network flow approach. In the experiments, we adopt the microarray dataset containing 62 prostate cancer samples and 41 normal samples, 108 known prostate cancer genes and 322 approved drug targets treated in human extracted from DrugBank database to be candidate proteins as our test data. Using our method, we prioritize the candidate proteins and validate them to the known prostate cancer drug targets. RESULTS: We successfully identify potential drug targets which are strongly related to the well known drugs for prostate cancer treatment and also discover more potential drug targets which raise the attention to biologists at present. We denote that it is hard to discover drug targets based only on differential expression changes due to the fact that those genes used to be drug targets may not always have significant expression changes. Comparing to previous methods that depend on the network topology attributes, they turn out that the genes having potential as drug targets are weakly correlated to critical points in a network. In comparison with previous methods, our results have highest mean average precision and also rank the position of the truly drug targets higher. It thereby verifies the effectiveness of our method. CONCLUSIONS: Our method does not know the real ideal routes in the disease network but it tries to find the feasible flow to give a strong influence to the disease genes through possible paths. We successfully formulate the identification of drug target prediction as a maximum flow problem on biological networks and discover potential drug targets in an accurate manner.