Microarray image compression: SLOCO and the effect of information loss

Microarray image technology is a powerful tool for monitoring the expression of thousands of genes simultaneously. Each microarray experiment produces immense amounts of image data, and efficient storage and transmission requires compression that utilizes the microarray image's structure and unique analysis goals. We have developed a progressive compression scheme for microarray images which can be either lossy or lossless. Our scheme has a coded data structure that allows fast decoding and reprocessing of image subsets, and includes summary statistics and image segmentation information. Since visual fidelity is not the end goal for microarray images, we introduce a new measure of distortion for lossy compression: the sensitivity of microarray information extraction to compression loss. We find that a lossy compression ratio of 8:1 for cDNA microarrays minimally affects downstream processing. The average lossless compression ratio is 1.83:1 for cDNA images and 2.43:1 for inkjet images, comparable to state-of-the-art lossless schemas, yet with added flexibility and information.

Electrocardiogram (ECG) compression can significantly reduce the storage and transmission burden in telemedicine applications. In this paper, an improved ECG data compression method using ASCII character encoding is proposed. A QRS detection algorithm firstly applied to separate QRS and non-QRS region. Then, the QRS regions of ECG signal are compressed using lossless method and non-QRS regions are compressed using lossy method, which improves the compression ratio and minimise the reconstruction error between original and reconstructed ECG signal. All tested ECG records are selected from the PTB database. Statistical results show that the compression performance of the proposed method is better than both lossless and lossy ECG data compression technique. The average compression ratio (CR) and percent root mean square difference (PRD) obtained using proposed ECG data compression method are 27.40 and 4.753 respectively, which is better than other existing ECG data compression method.

cDNA microarrays is one of the most fundamental and powerful tools in biotechnology. Despite its relatively late discovery in 1995, it has since been utilized in many biomedical applications such as cancer research, infectious disease diagnosis and treatment, toxicology research, pharmacology research, and agricultural development. The reason for its broad use is that it enables scientists to analyze simultaneously the expression levels of thousands of genes over different samples (Leung et al., 2003). More precisely, the process of a microarray experiment (Campbell et al., 2007) starts with the selection of a set of DNA probes that are of particular interest. A robot places the selected DNA probes on a glass slide, creating an invisible array of DNA dots. Two distinct populations of mRNAs (messenger RNAs) are then isolated from a control sample (i.e a cell developed under normal conditions) and a test sample (i.e. a cell developed under a specific treatment). The mRNA populations are reversely transcribed into cDNA (complementary DNA) populations which in turn are colored with separate fluorescent dyes of different wavelengths (i.e. Cy3 and Cy5). The dyed cDNA populations are mixed with purified water and the solution is placed on the glass slide in order for the cDNA populations to be hybridized with the slide’s DNA dots. Finally, the hybridized glass slide is fluorescently scanned twice; one scan for each dye’s wavelength. Hence, two digital images are produced, one for each population of mRNA. Each digital image contains a number of spots (corresponding to the DNA-cDNA dots) of various fluorescence intensities. Given that the intensity of each spot is proportional to the hybridization level of the cDNAs and the DNA dots, the gene expression information is obtained by analyzing the digital images. As stated by Yang et al (Yang et. al, 2002), the process of analyzing a microarray image can be divided into three main phases, namely: “Gridding”, “Spot-Segmentation” and “SpotIntensity extraction”. During the 1st phase, the microarray image is segmented into numerous compartments, each containing one individual spot and background. During the 2nd phase each compartment is individually segmented into a spot area and a background area, while during the 3rd phase the brightness of each spot is calculated. The expressionlevels of the genes in these spots are a direct result of their individual brightness. Amongst the stages of the microarray-image analysis, spot-segmentation remains the most challenging one. Ideally, the existing spots inside the microarray image are aligned in 2D array layouts. These ‘ideal spots’ also have a circular 2D shape with fixed diameters, while

<p>Microarray technology allows the simultaneous monitoring of thousands of genes. Based on the gene expression measurements, microarray technology have proven powerful in gene expression profiling for discovering new types of diseases and for predicting the type of a disease. Gridding, segmentation and intensity extraction are the three important steps in microarray image analysis. Clustering algorithms have been used for microarray image segmentation with an advantage that they are not restricted to a particular shape and size for the spots. Instead of using single feature clustering algorithm, this paper presents multiple feature clustering algorithm with three features for each pixel such as pixel intensity, distance from the center of the spot and median of surrounding pixels. In all the traditional clustering algorithms, number of clusters and initial centroids are randomly selected and often specified by the user. In this paper, a new algorithm based on empirical mode decomposition algorithm for the histogram of the input image will generate the number of clusters and initial centroids required for clustering. It overcomes the shortage of random initialization in traditional clustering and achieves high computational speed by reducing the number of iterations. The experimental results show that multiple feature Fuzzy C-means has segmented the microarray image more accurately than other algorithms.</p>

Microarray technology has become a power tool in the field of bioinformatics. It is used to measure gene expression levels and similar to any other image capturing processes is prone to noise. There are different kinds of noise, during preparation, hybridization and scanning in microarray images which usually are modeled by Gaussian noise. Since introduction of wavelets in 1970s, many more forms and extensions of this transform have been developed and used, such as stationary wavelet transform (SWT), complex wavelet transform (CWT), curvelet transform (CURV) and contourlet transform (CNT). By developing of more sparse transforms, it is important to have a perspective of how efficient the transforms are in different applications, such as microarray image analysis. In this paper, we compare the efficiency of common sparse transforms including ordinary discrete wavelet transform (DWT), SWT, CWT, CURV, CNT, Contourlet-SD decomposition, steerable pyramid (STP) and shearlet transform (SHR) for microarray image denoising. Therefore after converting microarray image into x-let transform, BayesShrink method, soft and hard thresholding are used to perform denoising of these images. Both local and general thresholds are calculated for each subband in order to evaluate the effect of incorporating intrascale dependency on top of sparsity property in statistical modeling of x-let's coefficients. Our simulation results show that CWT and SHR outperforms the others when using global thresholding and SWT is the preferred transform when using local thresholding. Although STP and SHR have better performance for some criteria like structural similarity (SSIM) index, but CWT is faster.

Current gene-expression microarrays carry enormous amounts of information. Compression is necessary for efficient distribution and storage. This paper examines JPEG2000 compression of cDNA microarray images and addresses the accuracy of classification and feature selection based on decompressed images. Among other options, we choose JPEG2000 because it is the latest international standard for image compression and offers lossy-to-lossless compression while achieving high lossless compression ratios on microarray images. The performance of JPEG2000 has been tested on three real data sets at different compression ratios, ranging from lossless to 45:1. The effects of JPEG2000 compression/decompression on differential expression detection and phenotype classification have been examined. There is less than a 4% change in differential detection at compression rates as high as 20:1, with detection accuracy suffering less than 2% for moderate to high intensity genes, and there is no significant effect on classification at rates as high as 35:1. The supplementary material is available at http://gsp.tamu.edu/web2/Compression.

We use both lossy and lossless compression methods to reduce large amounts of digital data, making it easier to store and manage. Lossy and lossless compression is frequently used in astronomy, to condense the large amount of data collected by instruments namely, the James Webb Space Telescope and the Hubble Space Telescope. Similarly to diffraction images, astronomical images have large areas of empty space since stellar objects are very bright, yet show very small in images, which after a while, takes up a needless amount of limited digital storage space. Lossy compression achieves higher compression ratios but sacrifices some data quality in the process, while lossless compression preserves the data’s original form, but results in less storage savings. This work, conducted at Brookhaven National Laboratory’s National Synchrotron Light Source-II at the AMX and FMX macromolecular beam lines, aimed to evaluate the effectiveness of the lossy compression method known as Wavelet Compression. Wavelet compression is a method of data compression that uses wavelet transforms to reduce the size of data while preserving its essential features. Crystals are repetitive molecular assemblies that create diffraction patterns when exposed to X-rays, while detectors capture the amplitude of diffraction pattern spots (reflections). Similar to astronomical images, diffraction images contain large areas of empty space. This unnecessary data can quickly consume limited digital storage. By studying and optimizing compression methods, we not only improve data management but also advancements in various scientific disciplines.

An automated method for gridding and clustering-based segmentation of cDNA microarray images

The Geostationary Imaging Fourier Transform Spectrometer (GIFTS), as part of NASA's New Millennium Program, is an advanced instrument to provide high-temporal-resolution measurements of atmospheric temperature and water vapor, which will greatly facilitate the detection of rapid atmospheric changes associated with destructive weather events, including tornadoes, severe thunderstorms, flash floods, and hurricanes. The Committee on Earth Science and Applications from Space under the National Academy of Sciences recommended that NASA and NOAA complete the fabrication, testing, and space qualification of the GIFTS instrument and that they support the international effort to launch GIFTS by 2008. Lossless data compression is critical for the overall success of the GIFTS experiment, or any other very high data rate experiment where the data is to be disseminated to the user community in real-time and archived for scientific studies and climate assessment. In general, lossless data compression is needed for high data rate hyperspectral sounding instruments such as GIFTS for (1) transmitting the data down to the ground within the bandwidth capabilities of the satellite transmitter and ground station receiving system, (2) compressing the data at the ground station for distribution to the user community (as is traditionally performed with GOES data via satellite rebroadcast), and (3) archival of the data without loss of any information content so that it can be used in scientific studies and climate assessment for many years after the date of the measurements. In this paper we study lossless compression of GIFTS data that has been collected as part of the calibration or ground based tests that were conducted in 2006. The predictive partitioned vector quantization (PPVQ) is investigated for higher lossless compression performance. PPVQ consists of linear prediction, channel partitioning and vector quantization. It yields an average compression ratio of 4.65 on the GIFTS test data, which significantly outperforms the standard compression methods such as JPEG-2000, JPEG-LS, and CCSDS IDC 9/7M & 5/3.

A spot-adaptive compound clustering-enhancement-segmentation (CES) scheme was developed for the quantification of gene expression levels in microarray images. The CES-scheme employed 1/griding, for locating spot-regions, 2/Fuzzy C-means clustering, for segmenting spots from background, 3/ background noise estimation and spot's center localization, 4/emphasizing of spot's outline by the CLAHE image enhancement technique, 5/segmentation by the SRG algorithm, using information from step 3, and 6/microarray spot intensity extraction. Extracted intensities by the CES-Scheme were compared against those obtained by the MAGIC TOOL's SRG. Kullback-Liebler metric's values for the CES-Scheme were on average double than MAGIC TOOL's, with differences ranging from 1.45bits to 2.77bits in 7 cDNA images. Coefficient-of-Variation results showed significantly higher reproducibility (p<0.001) for the CES-Scheme in quantifying gene expression levels. Processing times for 1024×1024 16-bit microarray images containing 6400 spots were 300 and 487 seconds for the CES-Scheme and MAGIC TOOL respectively.

Wavelets are defined as mathematical functions that segment the data into different frequency levels. We can easily capture the fine and coarse details of an image or signal referred to as a subband. And it also helps in subband thresholding to achieve good compression performance. In recent days in telemedicine services, the handling of medical images is prominently increasing and it leads to the demand for medical image compression. While compressing the medical images, we have to concentrate on the data that holds important information, and at the same time, it must retain the image quality. Near-Lossless compression plays an essential role to achieve a better compression ratio than lossy compression and provides better quality than lossless compression. In this paper, we analyzed the sub-banding of Discrete Wavelet Transform (DWT) using different types of wavelets and made an optimal selection of wavelets for subband thresholding to attain a good compression performance with an application to medical images. We used Set Partitioning In Hierarchical Trees (SPIHT) compression scheme to test the compression performance of different wavelets. The Peak Signal to Noise Ratio (PSNR), Bits Per Pixel (BPP), Compression Ratio, and percentage of number of zeros are used as metrics to assess the performance of all the selected wavelets. And to find out its efficiency in possessing the essential information of medical images, the subband of the selected wavelets is further utilized to devise the near-lossless compression scheme for medical images.

In an earlier publication we reported on the effectiveness of the Burrows-Wheeler transformation followed by inversion coder (BWIC) in the lossless compression of DNA microarray images where we obtained gains of average 6.5% over generic image compressors. In this work, we propose an enhancement of our previous technique by exploiting the bit distribution of images. Using a simple statistical test, we first decide if it will be gainful to split a 16-bit microarray image into two 8-bit images. In case of splitting, it turns out that the first 8-bit image is highly compressible and we use BWIC to compress it. The second 8-bit image most often contains noise and the bit distribution can become nearly random. We use the Wald-Wolfowitz runs test of randomness to decide whether to compress the second 8-bit image with BWIC or not at all since attempting to compress random data usually results in a larger file size. On select microarray images, by splitting a 16-bit microarray image into 8-bit pieces and selectively compressing the pieces with BWIC, we can achieve upward of 3% compression gain over our previous work.

DNA microarray images, which are the final results of a microarray experiment, are usually massive in size. We have designed and implemented an algorithm for lossy compression of microarray images. First, spots are automatically extracted into spot sub-images. Then, within each subimage, a circle is matched to each spot. This is done via spatial optimization techniques. After that, our circle to square transform, C2S, is applied to the sub-image and a square image representing its spot part is achieved. Information regarding the coordinates of the C2S transform is also included in the final compressed file. The background parts which contain no biological information are neglected. The square shaped images are then put together and coded by means of DCT transform, quantization, and entropy coding.

The transcriptome plays an important role in the life of a cell. Detailed analysis of the transcriptome enables interpretation of its structure and functionality. High throughput sequencing technology significantly enhanced the understanding of transcriptome activity. The RNA-sequencing process currently provides the most accurate estimation of gene expression levels. Moreover, RNA-seq allows detection of isoform structure and novel RNA types along with transcription process details such as strand-specificity and much more. The first chapter of this thesis describes the history of transcriptome exploration and effective methods of RNA-seq application. Nevertheless, all steps of RNA-seq process can produce a number of biases that influence the investigation results. Some typical errors appearing during ligation and amplification procedures might be present in any high throughput sequencing experiment, while other biases occur only in cDNA synthesis or are specific for transcriptome activity. Quality control of sequencing data is important to verify and correct the analysis results. The second chapter of this thesis is devoted to the explanation of these issues and introduces a novel tool, Qualimap 2. This instrument computes detailed statistics and presents a number of plots based on RNA-seq alignment and counts data processing. The generated results enable detection of problems that are specific to RNA-seq experiments. Notably, the tool supports analysis of multiple samples in various conditions. Qualimap 2 was faithfully compared to other available tools and demonstrated superior functionality in multi-sample quality control. Importantly, RNA-seq can be applied in a relatively novel research area: detection of chimeric transcripts and fusion genes occurring due to genomic rearrangement. Since fusions are related to cancer, their discovery is important not only for science, but also allows medical use of RNA-seq. The third chapter is devoted to the current status of this approach and illustrates a novel toolkit called InFusion, which provides a number of novelties in chimera discovery from RNA- seq data such as detection of fusions arising from the combination of a gene and an intronic or intergenic region. Moreover, strand-specificity of expressed fusion transcripts can be detected and reported. InFusion was compared in detail to a number of other existing tools based on simulated and real datasets and demonstrated higher precision and recall. Overall, RNA- sequencing technology goes further and more specialized analysis abilities are becoming available. New applications of RNA sequencing and future directions of research are discussed in the last chapter.

MIGNON is a workflow for the analysis of RNA-Seq experiments, which not only efficiently manages the estimation of gene expression levels from raw sequencing reads, but also calls genomic variants present in the transcripts analyzed. Moreover, this is the first workflow that provides a framework for the integration of transcriptomic and genomic data based on a mechanistic model of signaling pathway activities that allows a detailed biological interpretation of the results, including a comprehensive functional profiling of cell activity. MIGNON covers the whole process, from reads to signaling circuit activity estimations, using state-of-the-art tools, it is easy to use and it is deployable in different computational environments, allowing an optimized use of the resources available.