Cx31.1 acts as a tumour suppressor in non-small cell lung cancer (NSCLC) cell lines through inhibition of cell proliferation and metastasis.

Reduced connexin expression and loss of gap junction function is a characteristic of many cancers, including lung cancer. However, there are little reports about the relation between Cx31.1 and lung cancer. This study was conducted to investigate the effect of Cx31.1 on non-small cell lung cancer (NSCLC). We found that the Cx31.1 was down-regulated in NSCLC cell lines, and the expression levels were reversely related with their metastatic potential. We ectopically expressed Cx31.1 in H1299 NSCLC cell line to examine the influence of Cx31.1 overexpression. The results showed that overexpression of Cx31.1 in H1299 cells reduced cell proliferation, induced a delay in the G(1) phase, inhibited anchorage-independent growth and suppressed cell migration and invasion. The cell cycle delay and cell migration and invasion suppressive effects of Cx31.1 were partially reversed by siRNA targeting mRNA of Cx31.1. Moreover, xenografts of Cx31.1 overexpressing H1299 cells showed reduced tumourigenicity. These results suggested that Cx31.1 has tumour-suppressive properties. Further investigation indicated that cyclin D3 may be responsible for Cx31.1-induced G(1) phase delay. Importantly, Cx31.1 increased the expression of epithelial markers, such as cytokeratin 18, and decreased expression of mesenchymal markers, such as vimentin, indicating a Cx31.1-mediated partial shift from a mesenchymal towards an epithelial phenotype. We concluded that Cx31.1 inhibit the malignant properties of NSCLC cell lines, the mechanisms under this may include regulation of EMT.

Meta-analysis of three polymorphisms in the steroid-5-alpha-reductase, alpha polypeptide 2 gene (SRD5A2) and risk of prostate cancer.

The steroid-5-alpha-reductase, alpha polypeptide 2 (SRD5A2) gene plays a crucial role in androgen metabolism pathway in human prostate. It encodes SRD5A2 enzyme, which catalyses testosterone to dihydrotestosterone (DHT). DHT is the main active structure binding with androgen receptor (AR). After the activation of AR, it further regulates a series of target genes in androgen metabolism pathway. However, no clear consensus has been reached on the association between the SRD5A2 V89L, A49T and TA repeat polymorphisms and prostate cancer (PCa) risk. Thus, we performed a meta-analysis of 31 association studies with 14,726 PCa cases and 15,802 controls. We found no association between PCa and 89L compared with 89V allele [odds ratio (OR) = 1.02, 95% confidence interval (CI) 0.98-1.06, P(heterogeneity) = 0.44]. The 49T allele showed a significantly elevated effect on the high stage (Stages III-IV) of PCa risk both under the dominant genetic model (OR = 2.13, 95% CI 1.44-3.15, P(heterogeneity) = 0.65) and in the contrast T versus A allele (OR = 2.06, 95% CI 1.41-3.02, P(heterogeneity) = 0.69). There was a significantly decreased association between PCa and long TA repeat as compared versus short TA repeat (OR = 0.86, 95% CI 0.74-1.00, P(heterogeneity) = 0.79). No significant between-study heterogeneity was found in all subjects under four genetic models (dominant model, recessive model, allele comparison and homozygosity comparison) for these three polymorphisms, respectively, so the fixed effects model was used to pool the result. Our result indicated that carriers of 49T might improve the risk of PCa in higher stages (Stages III-IV), carriers of long TA repeat might decrease the risk of PCa and 89L may not be an important risk factor for PCa. However, due to the limited sample sizes, this meta-analysis did not achieve sufficiently conclusive results. Still more well-designed studies should be performed to clarify the role of these three polymorphisms in the development of PCa.

Cellular and molecular mechanisms of photodynamic hypericin therapy for nasopharyngeal carcinoma cells.

Hypericin-mediated photodynamic therapy (HY-PDT) has become a potential treatment for tumors and nonmalignant disorders. Some studies reported that HY-PDT could lead to apoptosis in some carcinoma cells. However, the molecular mechanism of HY-PDT remains unknown. In this study, we evaluated the molecular mechanisms of hypericin associated with light-emitting diode irradiation on the poorly differentiated human nasopharyngeal carcinoma cell line CNE-2 in vitro. To comprehensively understand the effects of HY-PDT on CNE-2 cells, we detected cell viability, cell cycle, apoptosis, intracellular glutathione content, and intracellular caspase (caspase-9, caspase-3, and caspase-8) activity. Furthermore, we performed genome-wide expression analysis via microarrays at different time points in response to HY-PDT, and we found that differentially expressed genes were highly enriched in the pathways related to reactive oxygen species generation, mitochondrial activity, DNA replication and repair, cell cycle/proliferation, and apoptosis. These results were consistent with our cytology test results and demonstrated that caspase-dependent apoptosis occurred after HY-PDT. Taken together, both cellular and molecular data revealed that HY-PDT could inhibit the growth of CNE-2 cells and induce their apoptosis.

Selection of reliable reference genes for gene expression study in nasopharyngeal carcinoma.

AIM: To construct a system for selecting reference genes (RGs) and to select the most optimal RGs for gene expression studies in nasopharyngeal carcinoma (NPC). METHODS: The total RNAs from 20 NPC samples were each labeled with Cy5-dUTP. To create a common control, the total RNA from 15 nasopharyngeal phlogistic (NP) tissues was mixed and labeled via reverse transcription with Cy3-dUTP. cDNA microarrays containing 14 112 genes were then performed. A mathematical approach was constructed to screen stably expressed genes from the microarray data. Using this method, three genes (YARS, EIF3S7, and PFDN1) were selected as candidate RGs. Furthermore, 7 commonly used RGs (HPRT1, GAPDH, TBP, ACTB, B2M, G6PDH, and HBB) were selected as additional potential RGs. Real-time PCR was used to detect these 10 candidate genes' expression levels and the geNorm program was used to find the optimal RGs for NPC studies. RESULTS: On the basis of the 10 candidate genes' expression stability level, geNorm analysis identified the optimal single RG (YARS or HPRT1) and the most suitable set of RGs (HPRT1, YARS, and EIF3S7) for NPC gene expression studies. In addition, this analysis determined that B2M, G6PDH, and HBB were not appropriate for use as RGs. Interestingly, ACTB was the least stable RG in our study, even though previous studies had indicated that it was one of the most stable RGs. Three novel candidate genes (YARS, EIF3S7, and PFDN1), which were selected from microarray data, were all identified as suitable RGs for NPC research. A RG-selecting system was then constructed, which combines microarray data analysis, a literature screen, real-time PCR, and bioinformatic analysis. CONCLUSION: We construct a RG-selecting system that helps find the optimal RGs. This process, applied to NPC research, determined the single RG (YARS or HPRT1) and the set of RGs (HPRT1, YARS, and EIF3S7) that are the most suitable internal controls.

Global gene expression analysis reveals reduced abundance of putative microRNA targets in human prostate tumours.

BACKGROUND: Recently, microRNAs (miRNAs) have taken centre stage in the field of human molecular oncology. Several studies have shown that miRNA profiling analyses offer new possibilities in cancer classification, diagnosis and prognosis. However, the function of miRNAs that are dysregulated in tumours remains largely a mystery. Global analysis of miRNA-target gene expression has helped illuminate the role of miRNAs in developmental gene expression programs, but such an approach has not been reported in cancer transcriptomics. RESULTS: In this study, we globally analysed the expression patterns of miRNA target genes in prostate cancer by using several public microarray datasets. Intriguingly, we found that, in contrast to global mRNA transcript levels, putative miRNA targets showed a reduced abundance in prostate tumours relative to benign prostate tissue. Additionally, the down-regulation of these miRNA targets positively correlated with the number of types of miRNA target-sites in the 3' untranslated regions of these targets. Further investigation revealed that the globally low expression was mainly driven by the targets of 36 specific miRNAs that were reported to be up-regulated in prostate cancer by a miRNA expression profiling study. We also found that the transcript levels of miRNA targets were lower in androgen-independent prostate cancer than in androgen-dependent prostate cancer. Moreover, when the global analysis was extended to four other cancers, significant differences in transcript levels between miRNA targets and total mRNA backgrounds were found. CONCLUSION: Global gene expression analysis, along with further investigation, suggests that miRNA targets have a significantly reduced transcript abundance in prostate cancer, when compared with the combined pool of all mRNAs. The abnormal expression pattern of miRNA targets in human cancer could be a common feature of the human cancer transcriptome. Our study may help to shed new light on the functional roles of miRNAs in cancer transcriptomics.

Generation and analysis of expressed sequence tags from a cDNA library of Moniezia expansa.

The Moniezia expansa is a parasite of sheep that causes diarrhea and fleshless leading to stockbreeding losses. A genomic resource for large-scale molecular studies in this parasite is lacking. To study the gene expression including development, digestion and reproduction organs of M. expansa, a cDNA library had been constructed and expressed sequence tags (ESTs) had been analyzed, which were helpful for the development of a powerful microarray platform which are used to analyze gene expression in this species. cDNAs are useful resources in annotating genes and providing functional analysis of genes. In this study, a cDNA library from adult cestode of M. expansa was created and 2642 ESTs from 5'-ends of the cDNA clones representing 1081 unigenes were obtained. Cluster analysis of these ESTs allowed identification of 1081 unique sequences containing 351 contigs and 730 singletons. BLASTX searches identified 780 significant (E-value<10(-5)) hits and further Gene Ontology (GO) analysis was used to annotate these genes. All the EST sequences were deposited under dbEST in GenBank (GenBank: FE905224-FE905315, FE942104-FE942773, FE969189-FE969190, FF677548-FF677734, FF848124-FF848253). Although we only obtained 1081 unigenes, the set of ESTs identified represents a significant proportion of the M. expansa and provides molecular resource for the development of microarrays for gene expression studies concerning development, metabolism and reproduction.

Simultaneously detection of genomic and expression alterations in prostate cancer using cDNA microarray.

BACKGROUND: Prostate cancer is a common disease among men but the knowledge of the prostate carcinogenesis is still limited. METHODS: cDNA microarray-based comparative genomic hybridization (CGH) and expression profiling were performed to screen the genomic and the expression changes in prostate cancer respectively. The two data were integrated to study the influence of genomic aberrations on gene expression and seek for the genes with their expression affected by the genomic aberrations. Real-time PCR was performed to evaluate the array data. RESULTS: Array-based CGH detected gains at 2q, 3p/q, 5q, 6q, 8q, 9p, 10p/q, 11q, 12p, 14q, and 19p/q and losses at 1p, 2p, 4q, 6p/q, 7p, 11p/q, 12q, 17p/q, 19p/q, and Xp/q in more than 20% prostate tumors and narrowed these aberrations. For example, the gain of 8q was mapped to five minimal regions. Novel aberrations were also identified, such as loss at Xq21.33-q22.2. Expression profiling discovered the significant biological processes involved in the prostate carcinogenesis, such as exogenous antigen presentation via MHC class II and protein ubiquitination. Integration analysis revealed a weak positive correlation between genomic copy number and gene expression level. Fifty-three genes showed their expression directly affected by the genomic aberrations possibly, including more than one member of Ras superfamily and major histocompatibility complex (MHC). These genes are involved in multiple biological processes. CONCLUSIONS: Integration of the CGH and expression data provided more information than separate analysis. Although the direct influence of genomic aberrations on gene expression seems weak, the influence can be extended by indirect regulation through a few directly affected genes. Because the influence can be persistent, the genes directly affected by the genomic aberrations may play key roles in the prostate carcinogenesis and are worth further analysis.

No association of ERCC1 C8092A and T19007C polymorphisms to cancer risk: a meta-analysis.

ERCC1 (excision repair cross complementation group 1) is a subunit of the nucleotide excision repair complex, which can perform DNA strand incision correction of DNA damage. Association studies on the ERCC1 polymorphisms (C8092A and T19007C) in cancer had shown conflicting results. We performed a meta-analysis from all eligible case-control studies to assess the purported associations. Overall, the 19007C allele (3 853 patients and 4 349 controls) showed no significant effect on cancer risk compared to 19007T allele (P=0.39, odds ratio (OR)=0.95; 95% confidence interval (CI) 0.85-1.06, P(heterogeneity)=0.001) in all subjects. Meta-analysis under other genetic contrasts did not reveal any significant association of T19007C to cancer in all subjects, Caucasians and Asians. The 19007C allele (2 279 patients and 2 808 controls) showed no significant effect on lung cancer risk compared to 19007T allele (P=0.72, OR=0.94, 95% CI 0.69-1.29, P(heterogeneity)=0.0001) in all subjects. No significant effect of 8092A allele (3 865 patients and 3 750 controls) on cancer risk in all subjects (P=0.85, OR=1.01, 95% CI 0.94-1.08, P(heterogeneity)=0.92) and in Caucasians and Asians compare to 8092C. No evidences of association of C8092A (501 patients and 620 controls) to squamous cell carcinoma were found. The accumulated evidence indicated ERCC1 T19007C and C8092A might not be risk factors for cancer. Significant between-study heterogeneity existed in T19007C, which arose from a study showing significant protecting effect of 19007C allele compare to 19007T allele in smokers. More studies based on larger, stratified case-control population should be required to further evaluate the role of ERCC1 C8092A and T19007C polymorphisms in different cancer, especially in smokers.

Hepatic differentiation and transcriptional profile of the mouse liver epithelial progenitor cells (LEPCs) under the induction of sodium butyrate.

The liver regenerates by progenitor cells when it is damaged in chronic liver diseases and extensive damage. The progenitor cells, also termed "oval cells" according to their morphological traits, can differentiate into hepatocytes and bile duct cells in vivo. To better understand the transcriptional pattern that accompanies the hepatic differentiation of oval cells, we applied cDNA microarray to analyze the oval cell-derived liver epithelial progenitor cells (LEPCs) during in vitro induced differentiation. Upon exposure to sodium butyrate, a histone deacetylase inhibitor, cultured LEPCs differentiate and express functional hepatocyte markers albumin, tryptophan 2, 3-dioxygenase and alcohol dehydrogenase. For expression profiling, cells were harvested at 6 h, 12 h, 1 d, 3 d and 7 d after exposure to sodium butyrate. After analyzing the microarray data by SOM clustering, total of 796 differentially regulated genes were grouped into 48 clusters. Consistent with the phenotype change of LEPCs after sodium butyrate treatment, many hepatocyte functional genes are revealed by analyzing the clusters containing genes up-regulated through all the time points. The clusters, containing down-regulated genes immediately after the induction, are also analyzed. The microarray data was validated by analyzing the expression of selected genes by quantitative real-time PCR. A set of genes expressed synergistically in these clusters may play a central role during the process of differentiation. Sodium butyrate decreases cyclin B1 and Cdk4 expression, which would be associated with LEPCs growth arrest shortly after treatment. Bmi1, a polycomb group protein, is also down-regulated immediately after treatment and remains at a low level during the induction. These findings highlight the key molecular mechanisms by which sodium butyrate, mediates its effects on cell growth arrest and induction of differentiation. In conclusion, our data reflect a global view of gene expression during hepatic differentiation of LEPCs induced by sodium butyrate.

Identification of genes differentially regulated by transcription factor, AP-2delta.

In an effort to better understand the transcriptional program regulated by transcription factor, AP-2delta, we used cDNA microarray to evaluate the relative expression of human genes in AD293 cells by exogenous expression of AP-2delta. Microarray showed 29 genes that were up-regulated and 39 genes with a down-regulated expression pattern. Among the identified genes were those encoding transcription factors, signal transduction molecules, kinases, as well as genes regulating cell growth, differentiation, and cell adhesion, a set of genes suggested to be regulated by AP-2. The results of microarray was validated for some regulated genes by real-time PCR analysis. These observations shed novel insight into the mechanism of AP-2delta action and provide a range of target genes for further investigation.

Cap1p plays regulation roles in redox, energy metabolism and substance transport: an investigation on Candida albicans under normal culture condition.

Cap1p, a transcription factor in Candida albicans, is believed to be required for tolerance to oxidative stress. However, no information is available concerning its function on basal transcriptional profile. In this study, differentially expressed genes between the CAP1-deleted strain and its parental strain under normal culture condition were identified through microarray analysis. Notably, among the 48 down-regulated genes with the deletion of CAP1, there were three clusters, functionally related to intracellular redox, energy metabolism and substance transport. IPF7817, IPF11105 and FDH11, the three putative Cap1p target genes functionally related to redox, were shown to be activated by oxidative stress in a Cap1p-dependent manner. Furthermore, rhodamine 6G efflux analyses demonstrated that Cap1p contributed to the energy-driven efflux. Taken together, these results reveal that Cap1p plays a significant role in redox status regulation, energy metabolism and substance transport under normal culture condition.

Event-specific detection of seven genetically modified soybean and maizes using multiplex-PCR coupled with oligonucleotide microarray.

With the increasing development of genetically modified organism (GMO) detection techniques, the polymerase chain reaction (PCR) technique has been the mainstay for GMO detection. An oligonucleotide microarray is a glass chip to the surface of which an array of oligonucleotides was fixed as spots, each containing numerous copies of a sequence-specific probe that is complementary to a gene of interest. So it is used to detect ten or more targets synchronously. In this research, an event-specific detection strategy based on the unique and specific integration junction sequences between the host plant genome DNA and the integrated gene is being developed for its high specificity using multiplex-PCR together with oligonucleotide microarray. A commercial GM soybean (GTS 40-3-2) and six GM maize events (MON810, MON863, Bt176, Bt11, GA21, and T25) were detected by this method. The results indicate that it is a suitable method for the identification of these GM soybean and maizes.

Cap1p is involved in multiple pathways of oxidative stress response in Candida albicans.

Cap1p, a transcription factor in Candida albicans, is thought to participate in oxidative stress tolerance, but the pathways involved are still unclear. The study was designed to reveal the possible pathways by examining changes in the transcription profile after H2O2 treatment with both the cap1-deleted strain CJD21 and its parental strain CAI4 using microarray analysis. Of the identified 89 genes differentially expressed in CAI4 after exposure to H2O2, 76 genes were in a Cap1p-dependent expression pattern. We have shown that Cap1p is involved in the oxidative stress response in C. albicans via multiple pathways, including the cellular antioxidant defense system (e.g., thioredoxin reductase, glutathione reductase, glutathione S-transferase), carbohydrate metabolism and energy metabolism (e.g., glucose-6-phosphate dehydrogenase, transaldolase, glyoxalase I, NADH-dependent flavin oxidoreductase), protein degradation (e.g., 26S proteasome regulatory subunit, ubiquitin-specific protease), ATP-dependent RNA helicase (e.g., DEAD box protein ATP-dependent RNA helicase), and resistance pathways (e.g., multidrug resistance protein, ABC transporter essential for cadmium resistance). Real-time reverse transcription-PCR analysis further confirmed the results of microarray. Collectively, this study provides new insight into the biological functions of Cap1p in oxidative stress response.

LS-CAP: an algorithm for identifying cytogenetic aberrations in hepatocellular carcinoma using microarray data.

Biological techniques such as Array-Comparative genomic hybridization (CGH), fluorescent in situ hybridization (FISH) and affymetrix single nucleotide pleomorphism (SNP) array have been used to detect cytogenetic aberrations. However, on genomic scale, these techniques are labor intensive and time consuming. Comparative genomic microarray analysis (CGMA) has been used to identify cytogenetic changes in hepatocellular carcinoma (HCC) using gene expression microarray data. However, CGMA algorithm can not give precise localization of aberrations, fails to identify small cytogenetic changes, and exhibits false negatives and positives. Locally un-weighted smoothing cytogenetic aberrations prediction (LS-CAP) based on local smoothing and binomial distribution can be expected to address these problems. LS-CAP algorithm was built and used on HCC microarray profiles. Eighteen cytogenetic abnormalities were identified, among them 5 were reported previously, and 12 were proven by CGH studies. LS-CAP effectively reduced the false negatives and positives, and precisely located small fragments with cytogenetic aberrations.

Genes differentially expressed in responsive and refractory acute leukemia.

DNA microarray in 22 patients with acute leukemia revealed genes which were differentially expressed. Ribosomal protein SA (RPSA), minichromosome maintenance deficient 2 (MCM2) and heterogeneous nuclear ribonucleoprotein A1 (HNRPA1) were significantly upregulated (p<0.05, t test) in refractory patients, suggesting that they may play a role in refractoriness in acute leukemia and could be biomarkers of prognosis.

Analysis of common gene expression patterns in four human tumor cell lines exposed to camptothecin using cDNA microarray: identification of topoisomerase-mediated DNA damage response pathways.

Camptothecin (CPT) is a potent inhibitor of DNA topoisomerase I with a wide spectrum of anti-tumor activity. Relatively little information is available regarding the relation of known topoisomerase-mediated DNA damage with other intracellular pathways. To gain an insight into the intracellular molecular mechanisms of Topoisomerase I inhibitor camptothecin-mediated DNA damage leading to cell death, we used a high-density cDNA microarray to assess sensitive early gene expression profiles in SGC7901 (gastric cancer), Hela (cervical adenocarcinoma), K562 (chronic myelogenous leukemia) and HL60 (promyelocytic leukemia) tumor cells stimulated with camptothecin for 1 h at the concentrations of GI50 (50 % growth inhibition after 24 h of treatment). Analysis of the differentially expressed genes obtained 29 response genes common to all four cell lines. Moreover, these cell lines also shared the direction of regulation. Most of these common response genes were functionally related to cell proliferation or apoptosis, and some of them were involved in ATM (ataxia-telangiectasia mutated) and ATR (ATM-and Rad3 related) checkpoint pathways, JNK (c-Jun N-terminal kinase) pathway, the survival phosphatidylinositol (PI) 3 kinase-Akt-dependent pathway, mitochondrial cell death pathway, endoplasmic reticulum (ER)-related cell death pathway, and to ubiquitin/proteasome dependent protein degradation pathway. The data provides evidence for a linkage between topoisomerase-mediated DNA damage and intracellular signaling events, which may facilitate our understanding of the camptothecin mediated molecular mechanisms of action.

Finding dominant sets in microarray data.

Clustering allows us to extract groups of genes that are tightly coexpressed from Microarray data. In this paper, a new method DSF_Clust is developed to find dominant sets (clusters). We have preformed DSF_Clust on several gene expression datasets and given the evaluation with some criteria. The results showed that this approach could cluster dominant sets of good quality compared to kmeans method. DSF_Clust deals with three issues that have bedeviled clustering, some dominant sets being statistically determined in a significance level, predefining cluster structure being not required, and the quality of a dominant set being ensured. We have also applied this approach to analyze published data of yeast cell cycle gene expression and found some biologically meaningful gene groups to be dug out. Furthermore, DSF_Clust is a potentially good tool to search for putative regulatory signals.

Analysis of the factors affecting the accuracy of detection for single base alterations by oligonucleotide microarray.

The oligonucleotide microarray, a high-throughput polymorphism detection technology, holds great promise for the characterization of complex genetic variance. To achieve greater sensitivity and specificity for it to be an effective platform technology we present results and discuss some of the factors influencing signal intensities and single-mismatch discrimination in array-based mutation/SNP detection. Probes with a series of concentrations were spotted onto the slide in order to find the optimal concentration with the identifiable satisfying signals and the stable ratios between matched and mismatched probes. It was found that under our experimental conditions, when the initial probe concentration is higher than the maximum immobilization capability of the slide (7.5 microM), the hybridization signal will be saturated and the ratio between matched and mismatched probes will be more stable than at a lower probe concentration. Considering the cost of probes and the systematic stability, a constant spotting concentration of 10 microM was selected. The stability of different types of mismatched oligo-DNA duplexes on the glass surface was also confirmed. The results show that the order of stability of mismatched oligo-DNA duplexes on a glass surface is in general agreement with previous reports conducted using liquid and polyacrylamide gel pads. This suggests that the influence of the mismatched base pair on the stability of the duplex in a solid hybridization system is similar to that in the solution hybridization environment.

Cloning and expression of human GTDC1 gene (glycosyltransferase-like domain containing 1) from human fetal library.

The glycosyltransferases (GTs) catalyze the synthesis of the carbohydrate portions of glycoproteins, glycolipids, and proteoglycans. Here we report the cloning and characterization of a novel human GTDC1 (glycosyltransferase-like domain containing 1) gene, which locates on human chromosome 2q22. The GTDC1 cDNA is 2954 bp in length, encoding a putative protein of 458 amino acids. At protein level human GTDC1 has 75 and 37% identity with its homologous counterparts in the mouse and fruitfly, respectively. RT-PCR analysis revealed its relatively high expression level in the adult lung, spleen, testis, and peripheral blood leukocyte.

Determining a detectable threshold of signal intensity in cDNA microarray based on accumulated distribution.

In microarray data mining, one of the key problems is how to handle weak signals. Based on a bent piecewise linear accumulated distribution generally found in the microarray data, a new detectable threshold finding method is proposed to filter genes with unreliable information in this paper. More reliable and reproducible data is produced for the subsequent data mining.