An algorithm for finding biologically significant features in microarray data based on a priori manifold learning

Identifying Significant Features in Cancer Methylation Data Using Gene Pathway Segmentation

In order to provide the most effective therapy for cancer, it is important to be able to diagnose whether a patient's cancer will respond to a proposed treatment. Methylation profiling could contain information from which such predictions could be made. Currently, hypothesis testing is used to determine whether possible biomarkers for cancer progression produce statistically significant results. However, this approach requires the identification of individual genes, or sets of genes, as candidate hypotheses, and with the increasing size of modern microarrays, this task is becoming progressively harder. Exhaustive testing of small sets of genes is computationally infeasible, and so hypothesis generation depends either on the use of established biological knowledge or on heuristic methods. As an alternative machine learning, methods can be used to identify groups of genes that are acting together within sets of cancer data and associate their behaviors with cancer progression. These methods have the advantage of being multivariate and unbiased but unfortunately also rapidly become computationally infeasible as the number of gene probes and datasets increases. To address this problem, we have investigated a way of utilizing prior knowledge to segment microarray datasets in such a way that machine learning can be used to identify candidate sets of genes for hypothesis testing. A methylation dataset is divided into subsets, where each subset contains only the probes that relate to a known gene pathway. Each of these pathway subsets is used independently for classification. The classification method is AdaBoost with decision trees as weak classifiers. Since each pathway subset contains a relatively small number of gene probes, it is possible to train and test its classification accuracy quickly and determine whether it has valuable diagnostic information. Finally, genes from successful pathway subsets can be combined to create a classifier of high accuracy.

Multiple Manifold Clustering Using Curvature Constrained Path

The problem of multiple surface clustering is a challenging task, particularly when the surfaces intersect. Available methods such as Isomap fail to capture the true shape of the surface near by the intersection and result in incorrect clustering. The Isomap algorithm uses shortest path between points. The main draw back of the shortest path algorithm is due to the lack of curvature constrained where causes to have a path between points on different surfaces. In this paper we tackle this problem by imposing a curvature constraint to the shortest path algorithm used in Isomap. The algorithm chooses several landmark nodes at random and then checks whether there is a curvature constrained path between each landmark node and every other node in the neighborhood graph. We build a binary feature vector for each point where each entry represents the connectivity of that point to a particular landmark. Then the binary feature vectors could be used as a input of conventional clustering algorithm such as hierarchical clustering. We apply our method to simulated and some real datasets and show, it performs comparably to the best methods such as K-manifold and spectral multi-manifold clustering.

A predictor for predicting Escherichia coli transcriptome and the effects of gene perturbations

Background

A means to predict the effects of gene over-expression, knockouts, and environmental stimuli in silico is useful for system biologists to develop and test hypotheses. Several studies had predicted the expression of all Escherichia coli genes from sequences and reported a correlation of 0.301 between predicted and actual expression. However, these do not allow biologists to study the effects of gene perturbations on the native transcriptome.

Results

We developed a predictor to predict transcriptome-scale gene expression from a small number (n = 59) of known gene expressions using gene co-expression network, which can be used to predict the effects of over-expressions and knockdowns on E. coli transcriptome. In terms of transcriptome prediction, our results show that the correlation between predicted and actual expression value is 0.467, which is similar to the microarray intra-array variation (p-value = 0.348), suggesting that intra-array variation accounts for a substantial portion of the transcriptome prediction error. In terms of predicting the effects of gene perturbation(s), our results suggest that the expression of 83% of the genes affected by perturbation can be predicted within 40% of error and the correlation between predicted and actual expression values among the affected genes to be 0.698. With the ability to predict the effects of gene perturbations, we demonstrated that our predictor has the potential to estimate the effects of varying gene expression level on the native transcriptome.

Conclusion

We present a potential means to predict an entire transcriptome and a tool to estimate the effects of gene perturbations for E. coli, which will aid biologists in hypothesis development. This study forms the baseline for future work in using gene co-expression network for gene expression prediction.