Iman Rezaeian

A fully automatic gridding method for cDNA microarray images

BackgroundProcessing cDNA microarray images is a crucial step in gene expression analysis, since any errors in early stages affect subsequent steps, leading to possibly erroneous biological conclusions. When processing the underlying images, accurately separating the sub-grids and spots is extremely important for subsequent steps that include segmentation, quantification, normalization and clustering.ResultsWe propose a parameterless and fully automatic approach that first detects the sub-grids given the entire microarray image, and then detects the locations of the spots in each sub-grid. The approach, first, detects and corrects rotations in the images by applying an affine transformation, followed by a polynomial-time optimal multi-level thresholding algorithm used to find the positions of the sub-grids in the image and the positions of the spots in each sub-grid. Additionally, a new validity index is proposed in order to find the correct number of sub-grids in the image, and the correct number of spots in each sub-grid. Moreover, a refinement procedure is used to correct possible misalignments and increase the accuracy of the method.ConclusionsExtensive experiments on real-life microarray images and a comparison to other methods show that the proposed method performs these tasks fully automatically and with a very high degree of accuracy. Moreover, unlike previous methods, the proposed approach can be used in various type of microarray images with different resolutions and spot sizes and does not need any parameter to be adjusted.

Identifying informative genes for prediction of breast cancer subtypes

It is known that breast cancer is not just one disease, but rather a collection of many different diseases occurring in one site that can be distinguished based in part on characteristic gene expression signatures. Appropriate diagnosis of the specific subtypes of this disease is critical for ensuring the best possible patient response to therapy. Currently, therapeutic direction is determined based on the expression of characteristic receptors; while cost effective, this method is not robust and is limited to predicting a small number of subtypes reliably. Using the original 5 subtypes of breast cancer we hypothesized that machine learning techniques would offer many benefits for feature selection. Unlike existing gene selection approaches, we propose a tree-based approach that conducts gene selection and builds the classifier simultaneously. We conducted computational experiments to select the minimal number of genes that would reliably predict a given subtype. Our results support that this modified approach to gene selection yields a small subset of genes that can predict subtypes with greater than 95% overall accuracy. In addition to providing a valuable list of targets for diagnostic purposes, the gene ontologies of selected genes suggest that these methods have isolated a number of potential genes involved in breast cancer biology, etiology and potentially novel therapeutics.

Identifying differentially expressed transcripts associated with prostate cancer progression using RNA-Seq and machine learning techniques

Background: Prostate cancer is complicated by a high level of unexplained variability in the aggressiveness of newly diagnosed disease. Given that this is one of the most prevalent cancers worldwide, finding biomarkers to effectively stratify high risk patient populations is a vital next step in improving survival rates and quality of life after treatment. Materials and Methods: In this study, we selected a dataset consisting of 106 prostate cancer samples, which represent various stages of prostate cancer and developed by RNA-Seq technology. Our objective is to identify differentially expressed transcripts associated with prostate cancer progression using pair-wise stage comparisons. Results: Using machine learning techniques, we identified 44 transcripts that are correlated to different stages of progression. Expression of an identified transcript, USP13, is reduced in stage T3 in comparison with stage T2c, a pattern also observed in breast cancer tumourigenesis. We also identified another differentially expressed transcript, PTGFR, which has also been reported to be involved in prostate cancer progression and has also been linked to breast, ovarian and renal cancers. Conclusions: The results support the use of RNA-Seq along with machine learning techniques as an essential tool in identifying potential biomarkers for prostate cancer progression. Further studies elucidating the biochemical role of identified transcripts in vitro are crucial in validating the use of these biomarkers in the prediction of disease progression and development of effective therapeutic strategies.

A novel model used to detect differential splice junctions as biomarkers in prostate cancer from RNA-Seq data

BACKGROUND In cancer alternative RNA splicing represents one mechanism for flexible gene regulation, whereby protein isoforms can be created to promote cell growth, division and survival. Detecting novel splice junctions in the cancer transcriptome may reveal pathways driving tumorigenic events. In this regard, RNA-Seq, a high-throughput sequencing technology, has expanded the study of cancer transcriptomics in the areas of gene expression, chimeric events and alternative splicing in search of novel biomarkers for the disease. RESULTS In this study, we propose a new two-dimensional peak finding method for detecting differential splice junctions in prostate cancer using RNA-Seq data. We have designed an integrative process that involves a new two-dimensional peak finding algorithm to combine junctions and then remove irrelevant introns across different samples within a population. We have also designed a scoring mechanism to select the most common junctions. CONCLUSIONS Our computational analysis on three independent datasets collected from patients diagnosed with prostate cancer reveals a small subset of junctions that may potentially serve as biomarkers for prostate cancer. AVAILABILITY The pipeline, along with their corresponding algorithms, are available upon request.

Obtaining biomarkers in cancer progression from outliers of time-series clusters

Studying the expression of transcripts throughout the various stages of prostate cancer may provide insight into the factors that influence the progression of the disease. Moreover, it may also reveal outlier transcripts, which have different trends than the majority of the transcripts. In this study, we use a time-series profile hierarchical clustering method to separate dissimilar groups of aligned transcripts that have maximum distance with the other group expression patterns throughout the various stages/sub-stages of prostate cancer progression. The isolated outliers can serve as biomarkers in analyzing different stages/sub-stages. This paper suggests that the combination of proper clustering, distance function and index validation for clusters are suitable model to find a pattern of trending for transcript abundance throughout different prostate cancer stages/sub-stages. The stages/sub-stages represent the time points, and the growth of the transcript abundance throughout those time points are cubic spline interpolated. The trending throughout those stages can lead to understanding the relationships among the transcripts and provide a better analysis of prostate cancer development through stages.

Breast cancer subtype identification using machine learning techniques

Breast cancer is the most commonly diagnosed cancer and the second leading cause of death among women worldwide. Accurate diagnosis of the specific subtypes of this disease is vital to ensure that patients are provided with the most effective therapeutic strategies that yield the greatest response. Using the newly proposed ten subtypes of breast cancer, we hypothesize that machine learning techniques offer many benefits for selecting the most informative biomarkers. Unlike existing gene selection approaches, in this study, a hierarchical classification approach is used that selects genes and builds the classifier concurrently. Our results support that this modified approach to gene selection yields a small subset of genes that can predict these ten subtypes with greater than 95% overall accuracy.

A new algorithm for finding enriched regions in ChIP-Seq data

Genome-wide profiling of DNA-binding proteins using ChIP-Seq has emerged as an alternative to ChIP-chip methods. Due to the large amounts of data produced by next generation sequencing, ChIP-Seq offers many advantages, such as much higher resolution, less noise and greater coverage than its predecessor, the ChIP-chip array. Multi-level thresholding algorithms have been applied to many problems in image and signal processing. These algorithms have been used for transcriptomics and genomics data analysis such as sub-grid and spot detection in DNA microarrays, and also for detecting significant regions based on next generation sequencing data. We show that our Optimal Multilevel Thresholding algorithm (OMT) has higher accuracy in detecting enriched regions (peaks) in comparison with previously proposed peak finders by testing three algorithms on the well-known FoxA1 Data set and also for four transcription factors (with a total of six antibodies) for Drosophila melanogaster. Using a small number of parameters is another advantage of the proposed method.

A parameterless automatic spot detection method for cDNA microarray images

Gridding cDNA microarray images is a critical step in gene expression analysis, since any errors in this stage are propagated in future steps in the analysis. We propose a fully automatic approach to detect the locations of the spots. The approach first detects and corrects rotations in the sub-grids by an affine transformation, followed by a polynomial-time optimal multi-level thresholding algorithm that finds the positions of the spots. Additionally, a new validity index is proposed in order to find the correct number of spots in each sub-grid, followed by a refinement procedure used to improve the performance of the method. Extensive experiments on real-life microarray images show that the proposed method performs these tasks automatically and with very high accuracy.

Finding Transcripts Associated with Prostate Cancer Gleason Stages Using Next Generation Sequencing and Machine Learning Techniques

Prostate cancer is a leading cause of death world-widely and the third leading cause of cancer death in Northen American men. Prostate cancer causes parts of the prostate cells to lose normal control of growth and division. The Gleason classification system is one of the known systems used to grade the aggressiveness of the prostate progression.

A new clustering method using wavelet based probability density functions for identifying patterns in time-series data

Clustering is a prominent method to identify similar patterns in large groups of data and can be beneficial in the bioinformatics studies due to this property. Classical methods such as k-means and maximum likelihood consider a mixture of Gaussian probability density function (PDF) of data and find clusters based on maximizing the PDF. However, correlation among different groups of data and existence of noise on the data make it difficult to correctly detect the correct number of clusters. Furthermore, the assumption of the Gaussian distance for the PDF is not necessarily true in real applications. This paper presents a new clustering method via wavelet-based probability density functions. For this purpose, first, a mixture of PDFs is estimated by the wavelet for each feature. After this, a multilevel thresholding method is implemented on the mixture of PDFs of each feature to obtain the clusters. Finally, a forward feature selection with memory is used to cluster the dataset based on combinations of the features. The profile alignment and agglomerative clustering (PAAC) index is applied for evaluating the number of clusters and features. Transcript expression throughout the various stages of prostate cancer is considered as a case study to identify patterns. The experimental results show the ability of the proposed method in detecting patterns of similar transcripts throughout disease progression. The results are promising in comparison with the other methods.