Detection of copy number variations based on a local distance using next-generation sequencing data.

As one of the main types of structural variation in the human genome, copy number variation (CNV) plays an important role in the occurrence and development of human cancers. Next-generation sequencing (NGS) technology can provide base-level resolution, which provides favorable conditions for the accurate detection of CNVs. However, it is still a very challenging task to accurately detect CNVs from cancer samples with different purity and low sequencing coverage. Local distance-based CNV detection (LDCNV), an innovative computational approach to predict CNVs using NGS data, is proposed in this work. LDCNV calculates the average distance between each read depth (RD) and its k nearest neighbors (KNNs) to define the distance of KNNs of each RD, and the average distance between the KNNs for each RD to define their internal distance. Based on the above definitions, a local distance score is constructed using the ratio between the distance of KNNs and the internal distance of KNNs for each RD. The local distance scores are used to fit a normal distribution to evaluate the significance level of each RDS, and then use the hypothesis test method to predict the CNVs. The performance of the proposed method is verified with simulated and real data and compared with several popular methods. The experimental results show that the proposed method is superior to various other techniques. Therefore, the proposed method can be helpful for cancer diagnosis and targeted drug development.

Integrating Somatic Mutations for Breast Cancer Survival Prediction Using Machine Learning Methods.

Breast cancer is the most common malignancy in women, and because it has a high mortality rate, it is urgent to develop computational methods to increase the accuracy of breast cancer survival predictive models. Although multi-omics data such as gene expression have been extensively used in recent studies, the accurate prognosis of breast cancer remains a challenge. Somatic mutations are another important and promising data source for studying cancer development, and its effect on the prognosis of breast cancer remains to be further explored. Meanwhile, these omics datasets are high-dimensional and redundant. Therefore, we adopted multiple kernel learning (MKL) to efficiently integrate somatic mutation to currently molecular data including gene expression, copy number variation (CNV), methylation, and protein expression data for the prediction of breast cancer survival. Before integration, the maximum relevance minimum redundancy (mRMR) feature selection method was utilized to select features that present high relevance to survival and low redundancy among themselves for each type of data. The experimental results demonstrated that the proposed method achieved the most optimal performance and there was a remarkable improvement in the prediction performance when somatic mutations were included, indicating that somatic mutations are critical for improving breast cancer survival predictions. Moreover, mRMR was superior to other feature selection methods used in previous studies. Furthermore, MKL outperformed the other traditional classifiers in multi-omics data integration. Our analysis indicated that through employing promising omics data such as somatic mutations and harnessing the power of proper feature selection methods and effective integration frameworks, the breast cancer survival predictive accuracy can be further increased, thereby providing a more optimal clinical diagnosis and more effective treatment for breast cancer patients.

Stratification of Breast Cancer by Integrating Gene Expression Data and Clinical Variables.

Breast cancer is a heterogeneous disease. Although gene expression profiling has led to the definition of several subtypes of breast cancer, the precise discovery of the subtypes remains a challenge. Clinical data is another promising source. In this study, clinical variables are utilized and integrated to gene expressions for the stratification of breast cancer. We adopt two phases: gene selection and clustering, where the integration is in the gene selection phase; only genes whose expressions are most relevant to each clinical variable and least redundant among themselves are selected for further clustering. In practice, we simply utilize maximum relevance minimum redundancy (mRMR) for gene selection and k-means for clustering. We compare the results of our method with those of two commonly used only expression-based breast cancer stratification methods: prediction analysis of microarray 50 (PAM50) and highest variability (HV). The result is that our method outperforms them in identifying subtypes significantly associated with five-year survival and recurrence time. Specifically, our method identified recurrence-associated breast cancer subtypes that were not identified by PAM50 and HV. Additionally, our analysis discovered three survival-associated luminal-A subgroups and two survival-associated luminal-B subgroups. The study indicates that screening clinically relevant gene expressions yields improved breast cancer stratification.

Author Correction: Niche harmony search algorithm for detecting complex disease associated high-order SNP combinations.

A correction to this article has been published and is linked from the HTML and PDF versions of this paper. The error has not been fixed in the paper.

Niche harmony search algorithm for detecting complex disease associated high-order SNP combinations.

Genome-wide association study is especially challenging in detecting high-order disease-causing models due to model diversity, possible low or even no marginal effect of the model, and extraordinary search and computations. In this paper, we propose a niche harmony search algorithm where joint entropy is utilized as a heuristic factor to guide the search for low or no marginal effect model, and two computationally lightweight scores are selected to evaluate and adapt to diverse of disease models. In order to obtain all possible suspected pathogenic models, niche technique merges with HS, which serves as a taboo region to avoid HS trapping into local search. From the resultant set of candidate SNP-combinations, we use G-test statistic for testing true positives. Experiments were performed on twenty typical simulation datasets in which 12 models are with marginal effect and eight ones are with no marginal effect. Our results indicate that the proposed algorithm has very high detection power for searching suspected disease models in the first stage and it is superior to some typical existing approaches in both detection power and CPU runtime for all these datasets. Application to age-related macular degeneration (AMD) demonstrates our method is promising in detecting high-order disease-causing models.

Network based stratification of major cancers by integrating somatic mutation and gene expression data.

The stratification of cancer into subtypes that are significantly associated with clinical outcomes is beneficial for targeted prognosis and treatment. In this study, we integrated somatic mutation and gene expression data to identify clusters of patients. In contrast to previous studies, we constructed cancer-type-specific significant co-expression networks (SCNs) rather than using a fixed gene network across all cancers, such as the network-based stratification (NBS) method, which ignores cancer heterogeneity. For each type of cancer, the gene expression data were used to construct the SCN network, while the gene somatic mutation data were mapped onto the network, propagated, and used for further clustering. For the clustering, we adopted an improved network-regularized non-negative matrix factorization (netNMF) (netNMF_HC) for a more precise classification. We applied our method to various datasets, including ovarian cancer (OV), lung adenocarcinoma (LUAD) and uterine corpus endometrial carcinoma (UCEC) cohorts derived from the TCGA (The Cancer Genome Atlas) project. Based on the results, we evaluated the performance of our method to identify survival-relevant subtypes and further compared it to the NBS method, which adopts priori networks and netNMF algorithm. The proposed algorithm outperformed the NBS method in identifying informative cancer subtypes that were significantly associated with clinical outcomes in most cancer types we studied. In particular, our method identified survival-associated UCEC subtypes that were not identified by the NBS method. Our analysis indicated valid subtyping of patient could be applied by mutation data with cancer-type-specific SCNs and netNMF_HC for individual cancers because of specific cancer co-expression patterns and more precise clustering.