Raw Array Research Articles

Machine learning-assisted sensing array for simultaneous discrimination of hypochlorite and hydroxyl radicals

Modern data analysis methods, including machine learning (ML), have become essential instruments for uncovering concealed patterns and deducing correlations that conventional analytical techniques may encounter limitations in handling. This study provides a detailed introduction to a three-dimensional sensor array for hypochlorite and hydroxyl radical developed using machine learning methods, based on the sensitive differentiation of spectroscopic properties using a single fluorescent probe. To quantitatively analyze hypochlorite and selectively identify hydroxyl radicals, the raw array signals were transformed into a synthetic index through principal component analysis (PCA). Subsequently, hierarchical clustering analysis (HCA) and density clustering analysis (DCA) were employed to establish a classifier capable of correlating data structure and similarity across various samples. The innovative approach demonstrated here not only facilitates rapid and precise identification of hypochlorite and hydroxyl radical species but also lays a sturdy foundation for accurately discerning environmental pollutants exhibiting similar properties. Moreover, by aligning with the overarching goal of cleaner production, this work contributes to the ongoing efforts aimed at mitigating environmental impacts and fostering a greener future.

Fast Supraharmonic Estimation Algorithm Based on Simplified Compressed Sensing Model

The computational time of compressed sensing algorithms applied to supraharmonic needs to be improved in online applications. In this paper, a simplified supraharmonic compressive sensing model is proposed. The model first detects the supraharmonic raw spectral array to obtain the estimated sparsity and the index of supraharmonic emissions, which simplifies the sensing matrix in the iteration according to the index and then shortens the whole iteration time of compressed sensing. The simulation verifies that the model can reduce the computation time to less than half of the original compressed sensing model and does not affect the computation accuracy. Finally, the online application effect of the algorithm is verified by experiments.

An Automated Approach to Assess Relative Galectin-Glycan Affinity Following Glycan Microarray Analysis.

Numerous studies have highlighted the utility of glycan microarray analysis for the elucidation of protein-glycan interactions. However, most current glycan microarray studies analyze glycan binding protein (GBP)-glycan interactions at a single protein concentration. While this approach provides useful information related to a GBP’s overall binding capabilities, extrapolation of true glycan binding preferences using this method fails to account for printing variations or other factors that may confound relative binding. To overcome this limitation, we examined glycan array binding of three galectins over a range of concentrations to allow for a more complete assessment of binding preferences. This approach produced a richer data set than single concentration analysis and provided more accurate identification of true glycan binding preferences. However, while this approach can be highly informative, currently available data analysis approaches make it impractical to perform binding isotherms for each glycan present on currently available platforms following GBP evaluation. To overcome this limitation, we developed a method to directly optimize the efficiency of assessing association constants following multi-GBP concentration glycan array analysis. To this end, we developed programs that automatically analyze raw array data (kdMining) to generate output graphics (kaPlotting) following array analysis at multiple doses. These automatic programing methods reduced processing time from 32.8 h to 1.67 min. Taken together, these results demonstrate an effective approach to glycan array analysis that provides improved detail and efficiency when compared to previous methods.

P-552 Analysis of the genomewide BAF profiles of selected SNPs allows reliable aneuploidy detection in preimplantation embryos, independent of haplotyping

Abstract Study question Can we reliably detect aneuploidy in trophectoderm samples using SNP array data, independent of haplotyping? Summary answer We identified all chromosomes with meiotic copy number gains and non-mosaic copy number losses larger than 5Mb that had been detected with Karyomapping. What is known already Currently, generic methods such as Karyomapping (Vitrolife) and OnePGT (Agilent), are commonly used for preimplantation genetic testing of monogenic disorders (PGT-M). The genome-wide approach has the advantage that also aneuploidy can be detected by visualizing the haplotypes, the raw B-allele frequency (BAF) and the copynumber (CN) or Log2 ratio (Log2R) values. However, these haplotyping methods have considerable shortcomings for PGT-A including the need of reference DNA samples from both sides of the family and the long time required to access all haplotype information. We aimed to circumvent these shortcomings with our own proprietary method. Study design, size, duration We retrospectively re-analyzed the available raw SNP array data from 359 embryos with trophectoderm biopsy between September 1 2015 and December 31 2017 that were diagnosed with Karyomapping (Vitrolife) as being not genetically transferable. The local ethical committee approved this study under number B.U.N. 143201731745. Written informed consent was available. We intended to identify anomalies >5Mb that are detectable by haplotyping. Participants/materials, setting, methods SNPs were categorized based on the parental SNP genotypes. Next, the BAF values for each category of SNPs (cBAF) were extracted from the raw SNP array data and visualized on genomewide plots. The obtained cBAF plots were analyzed for the presence of aneuploidy together with raw BAF and Log2R SNP data. The results from the interpretation of these profiles were compared with the findings previously obtained with Karyomapping (Log2R, raw BAF and haplotype profiles). Main results and the role of chance A low level maternal contamination was observed in 5 embryos by analysis of cBAF profiles that had not been detected using Karyomapping. In five other embryos, an abnormal ploidy was observed with both methods. In the remaining 8027 chromosomes from 349 embryos, all 70 chromosomes with both parental homolog (BPH) anomalies larger than 5Mb detected with Karyomapping were identified by analysis of cBAF profiles. In 68 chromosomes the type (segmental or whole chromosome) and the copy number (+1 or + 2 copies) of the detected BPH anomaly was identical with both methods. All 93 chromosomes with a copy number loss larger than 5Mb detected with Karyomapping were identified by analysis of cBAF profiles. For 92 out of 93 chromosomes the type of non-mosaic copy number loss was identical (segmental or whole chromosome). Out of 17 single parental homolog (SPH) copy number gains interpreted as non-mosaic by Karyomapping, 16 were interpreted as non-mosaic based on cBAF profiles. In one instance, the SPH copy number detected with Karyomapping was interpreted as a mosaic SPH trisomy using cBAF. Two chromosomes with mosaic anomalies and one chromosome without diagnosis with Karyomapping were interpreted as non-mosaic anomalies using cBAF. Limitations, reasons for caution In case of consanguinity aneuploidy may be more difficult to detect. The limit of detection for the size of segmental anomalies remains to be determined. Wider implications of the findings By analyzing the cBAF profiles we were able to detect all types of chromosome anomaly that are detectable by haplotyping. We were able to identify meiotic BPH copy number gains and determine the parent of origin for all anomalies without haplotyping and hence without the need for DNA from familymembers. Trial registration number not applicable

Strain-based delamination prediction in fatigue loaded CFRP coupon specimens by deep learning and static loading data

Machine learning (ML) methods for the structural health monitoring (SHM) of composite structures rely on sufficient domain knowledge as they typically demand to extract damage-sensitive features from raw data before training the ML model. In practice, prior knowledge is not available in most cases. Deep learning (DL) methods, on the other hand, can obtain higher-level features from raw input data and have proven superior in several applications. This paper proposes a Convolutional Neural Network (CNN) based approach for the delamination prediction in CFRP double cantilever beam (DCB) specimens using raw local array strain measurements via distributed optical fiber sensors. The conventional CNN architecture is modified to perform regression, as the delamination size is a continuous value. 1D and 2D CNN architectures are deployed and compared and different techniques are exploited to encode 1D spatial strain pattern series as 2D images. Raw strain patterns collected during static testing are used to train the CNNs, while testing is performed on unseen raw fatigue strain patterns, showing the CNN ability to automatically extract discriminative features from the non-pre-processed static strain pattern-based signals that generalize to raw fatigue signals as well. This strategy has the potential to reduce fatigue testing expenditures while also shortening the time required to gather training data.

Multidimensional Multiconvolution-Based Feature Extraction Approach for Drift Tolerant Robust Classifier for Gases/Odors

Recently, convolutional neural networks (CNNs) have been used for the classification of gases/odors. These methods involve statistical algorithms for drift compensation of sensor characteristics inhibiting its application in real time. In this letter, we have proposed a hybrid CNN model called “drift tolerant robust classifier (DTRC),” which extracts multidimensional features from the raw sensor array responses and automatically compensates for any drift in the sensor response characteristics. The proposed DTRC comprises of 1-D, 2-D, and 3-D convolutional layers in a hybrid manner to compensate for the referred drift without any statistical algorithm. The efficacy of DTRC has been evaluated on a popular dataset and its published results, which comprise of ten batches of sensor characteristics exhibiting drift over a period of three years. Our proposed DTRC outperformed the referred results. In another experiment, DTRC outperformed other state-of-the-art methods. The proposed CNN architecture (DTRC) is a simpler, lightweight CNN with multidimensional multiconvolution end-to-end architecture, suitable for real-time applications.

Lung cancer detection via breath by electronic nose enhanced with a sparse group feature selection approach

In the diagnosis of lung cancer, electronic nose (E-nose) has attracted much attention by detecting the volatile organic compounds (VOCs) in exhaled breath. In this research, an E-nose platform with a novel thermal desorption preconcentration subsystem is designed to verify whether analyzing VOCs can reliably differentiate lung cancer patients from healthy individuals and patients with benign pulmonary diseases. To this end, total 87 subjects (46 patients with lung cancer, 36 healthy volunteers and 5 patients with benign pulmonary diseases) are enrolled for the sensor array data collection. 13 composite features are extracted from each sensor, and some classical classifiers are established to demonstrate the feasibility of identifying lung cancer patients through VOCs. To improve the performance, considering the inherent characteristics of E-nose data, a sparse group feature selection (FS) method is applied to the raw sensor array data. It is observed that the FS method can reduce data dimensionality and improve the classification performance significantly. Statistical analysis for the effect of age and smoking habit on classification results shows that no significant influences are found.

Effective nested Kalman fusion for improving microelectromechanical system array performance

Microelectromechanical systems (MEMS) are widely used in the navigation field due to their low cost and easy integration. Its low positioning accuracy restricts its expansion into the high-end navigation field. To improve the performance of MEMS inertial devices, this paper proposes a nested Kalman fusion (NKF) for MEMS gyroscope array data fusion applied to the virtual gyroscope. First, the algorithm processes the raw gyroscope array data through Kalman filtering. Secondly, the obtained filtered array data converge as a virtual gyroscope by support degree data fusion—the NKF experimental data collected by the actual test. The experimental results show the zero-bias instability, angular random walk and rate ramp of the original data are improved by 10.64 dB, 12.45 dB and 10.26 dB, respectively, by the NKF algorithm. NKF can adjust the gyro parameters by about 6 dB in comparison with existing MEMS optimization algorithms.

Abstract 1202: Genetically inferred telomere length is associated with clonal copy number alterations in peripheral leukocytes

Abstract Telomeres are DNA-protein structures at the ends of chromosomes essential in maintaining chromosomal stability. Telomeres shorten with each cell division due to incomplete replication of the 3' DNA end resulting in age-related telomere attrition. Very short telomeres induce cellular senescence and/or apoptosis, but cancer cells can bypass this mechanism by upregulating telomerase, TP53 and/or RB. Recent observational studies have identified associations between longer telomeres and elevated cancer risk, including hematologic malignancies; but biologic mechanisms relating telomere length to cancer etiology remain unclear. Our study sought to better understand the links between telomere length and cancer risk by evaluating relationships between telomere length and (1) the acquisition of somatic copy number alterations (SCNAs), (2) the chromosomal distribution and copy number state of acquired SCNAs, and (3) the cellular fraction of clones carrying SCNAs. Genomic data were derived from 431,507 participants in the UK Biobank with genotyping array data (Affymetrix UK BiLEVE Axiom and Biobank Axiom arrays). Telomere length was genetically inferred using a polygenic risk score based on effect estimates and allelic counts of 20 common variants associated with leukocyte telomere length and subsequently standardized. SCNAs were called using raw genotyping array intensity data to detect allelic imbalances (BAF) and deviations from expected intensity (log R ratio) along with data on haplotype imbalances as implemented in MoChA. In total, 15,236 (3.5%) of individuals had a detectable clonal SCNA on an autosomal chromosome. Overall, longer genetically inferred telomere length was positively associated with the presence of an autosomal SCNA (OR=1.07, 95% CI=1.05-1.09, P=1.10 × 10−15). We noted no differences in estimates among strata of chromosomal event location (e.g., telomeric ends, interstitial or whole chromosome event; P= 0.34), suggesting no preference in the type of event associated with inferred telomere length. Likewise, event copy number status demonstrated similar effect sizes across strata of copy number state (e.g., gain, loss, neutral, and unknown events; P= 0.05). When investigating the relationship between inferred telomere length and proportion of cells affected by mosaic SCNAs, longer genetically inferred telomere length was associated with a higher cellular fraction of clones carrying SCNAs (β=0.004, 95% CI= 0.002-0.007, P= 5.60 × 10−4). Cumulatively, our population-based examination of inferred telomere length and SCNAs suggests inherited components of telomere length do not preferentially impact SCNA event location or copy number status, but rather are associated with cellular replicative potential possibly by promoting clonal expansion of leukocytes harboring SCNAs. Citation Format: Derek W. Brown, Shu-Hong Lin, Stephen J. Chanock, Sharon A. Savage, Mitchell J. Machiela. Genetically inferred telomere length is associated with clonal copy number alterations in peripheral leukocytes [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1202.

Преподаване на C++ като първи език за програмиране

Nowadays, most C++ introductory courses begin by studying the procedural constructs of C language, following the imperative-first strategy described in CC2001 [7]. While this has been a successful approach in the near past, our experience shows that it does not lead to the development of skills to create a good C++ code. That’s why in recent years we’ve been teaching C++ to beginners in a totally different way – with avoidance of C-style strings and raw arrays; delayed introduction to pointers (just after references); polymorphism with references instead of pointers; smart pointers instead of raw pointers; early use of standard library features, and writing modern C++ from day one.

Reliability of NAND Flash Arrays: A Review of What the 2-D–to–3-D Transition Meant

This paper reviews what changed in the reliability of NAND Flash memory arrays after the paradigm shift in technology evolution determined by the transition from 2-D to 3-D integration schemes. Starting from a quick glance at the fundamentals of raw array reliability, the reasons for its worsening with the evolution of 2-D technologies will be discussed, focusing on the physical phenomena which contributed more to that outcome. By exploring the dependence of the magnitude of these phenomena on cell and array parameters, the abrupt improvements achieved from the 3-D transition in terms of raw array reliability will then be explained, highlighting also that these improvements were turned into new opportunities for the technology. Finally, the physical issues specific to 3-D arrays will be addressed, providing a glimpse of the challenges that the NAND Flash technology will have to face from the standpoint of array reliability in the near future.

Computational and Statistical Analysis of Array-Based DNA Methylation Data.

The characterization of aberrant DNA methylation is emerging as a key part of the study of cancer development and phenotype. The technical advancements and decreasing costs of methods for high-throughput profiling of DNA methylation have brought about a high interest in the use of such methods in disease association studies. Here we discuss the principles for DNA methylation analysis using data from the Infinium DNA methylation BeadChip assays and describe the computational steps and statistical considerations going from processing of the raw array data to analysis of differential methylation. Moreover, we provide detailed guidelines on how to perform tumor subtype classification based on DNA methylation signatures.

MethCNA: a database for integrating genomic and epigenomic data in human cancer

BackgroundThe integration of DNA methylation and copy number alteration data promises to provide valuable insight into the underlying molecular mechanisms responsible for cancer initiation and progression. However, the generation and processing of these datasets are costly and time-consuming if carried out separately. The Illumina Infinium HumanMethylation450 BeadChip, initially designed for the evaluation of DNA methylation levels, allows copy number variant calling using bioinformatics tools.ResultsA substantial amount of Infinium HumanMethylation450 data across various cancer types has been accumulated in recent years and is a valuable resource for large-scale data analysis. Here we present MethCNA, a comprehensive database for genomic and epigenomic data integration in human cancer. In the current release, MethCNA contains about 10,000 tumor samples representing 37 cancer types. All raw array data were collected from The Cancer Genome Atlas and NCBI Gene Expression Omnibus database and analyzed using a pipeline that integrated multiple computational resources and tools. The normalized copy number aberration data and DNA methylation alterations were obtained. We provide a user-friendly web-interface for data mining and visualization.ConclusionsThe Illumina Infinium HumanMethylation450 BeadChip enables the interrogation and integration of both genomic and epigenomic data from exactly the same DNA specimen, and thus can aid in distinguishing driver from passenger mutations in cancer. We expect MethCNA will enable researchers to explore DNA methylation and copy number alteration patterns, identify key oncogenic drivers in cancer, and assist in the development of targeted therapies. MethCNA is publicly available online at http://cgma.scu.edu.cn/MethCNA.

Efficiency of different strategies to mitigate ascertainment bias when using SNP panels in diversity studies

BackgroundSingle nucleotide polymorphism (SNP) panels have been widely used to study genomic variations within and between populations. Methods of SNP discovery have been a matter of debate for their potential of introducing ascertainment bias, and genetic diversity results obtained from the SNP genotype data can be misleading. We used a total of 42 chicken populations where both individual genotyped array data and pool whole genome resequencing (WGS) data were available. We compared allele frequency distributions and genetic diversity measures (expected heterozygosity (He), fixation index (FST) values, genetic distances and principal components analysis (PCA)) between the two data types. With the array data, we applied different filtering options (SNPs polymorphic in samples of two Gallus gallus wild populations, linkage disequilibrium (LD) based pruning and minor allele frequency (MAF) filtering, and combinations thereof) to assess their potential to mitigate the ascertainment bias.ResultsRare SNPs were underrepresented in the array data. Array data consistently overestimated He compared to WGS data, however, with a similar ranking of the breeds, as demonstrated by Spearman’s rank correlations ranging between 0.956 and 0.985. LD based pruning resulted in a reduced overestimation of He compared to the other filters and slightly improved the relationship with the WGS results. The raw array data and those with polymorphic SNPs in the wild samples underestimated pairwise FST values between breeds which had low FST (<0.15) in the WGS, and overestimated this parameter for high WGS FST (>0.15). LD based pruned data underestimated FST in a consistent manner. The genetic distance matrix from LD pruned data was more closely related to that of WGS than the other array versions. PCA was rather robust in all array versions, since the population structure on the PCA plot was generally well captured in comparison to the WGS data.ConclusionsAmong the tested filtering strategies, LD based pruning was found to account for the effects of ascertainment bias in the relatively best way, producing results which are most comparable to those obtained from WGS data and therefore is recommended for practical use.

Light Field Compression on Sliced Lenslet Array

Recent advances in light field technologies is fostering the research and development of novel imaging applications. Such applications perform processing of the light field information to create new visual effects such as, for example, refocusing, perspective change, colour adjustment. Light field imaging is very data intensive compared with usual digital photographic imaging, and novel compression algorithms are needed for addressing the problem of light field storage and transmission. Raw digital light field images exhibit low spatial correlation if compared to regular images and hence the performance in terms of rate-distortion of current state-of-the-art image encoders can be superseded by devising novel image compression architectures. In this paper, an architecture for lossy compression of unfocused light field images is proposed. Raw light fields are preprocessed by demosaicing, devignetting and slicing of the raw lenslet array image. The slices are then compressed with the JPEG 2000 image coding standard. The performance of the proposed method is compared against direct application of JPEG 2000 compression on the 4D light field. The experimental analysis has been conducted under a set of different compression ratios and the obtained results show that the proposed method outperforms direct application of the reference architecture.

Light field compression on sliced lenslet array

Recent advances in light field technologies is fostering the research and development of novel imaging applications. Such applications perform processing of the light field information to create new visual effects such as, for example, refocusing, perspective change, colour adjustment. Light field imaging is very data intensive compared with usual digital photographic imaging, and novel compression algorithms are needed for addressing the problem of light field storage and transmission. Raw digital light field images exhibit low spatial correlation if compared to regular images and hence the performance in terms of rate-distortion of current state-of-the-art image encoders can be superseded by devising novel image compression architectures. In this paper, an architecture for lossy compression of unfocused light field images is proposed. Raw light fields are preprocessed by demosaicing, devignetting and slicing of the raw lenslet array image. The slices are then compressed with the JPEG 2000 image coding standard. The performance of the proposed method is compared against direct application of JPEG 2000 compression on the 4D light field. The experimental analysis has been conducted under a set of different compression ratios and the obtained results show that the proposed method outperforms direct application of the reference architecture.

SNPConvert: SNP Array Standardization and Integration in Livestock Species.

One of the main advantages of single nucleotide polymorphism (SNP) array technology is providing genotype calls for a specific number of SNP markers at a relatively low cost. Since its first application in animal genetics, the number of available SNP arrays for each species has been constantly increasing. However, conversely to that observed in whole genome sequence data analysis, SNP array data does not have a common set of file formats or coding conventions for allele calling. Therefore, the standardization and integration of SNP array data from multiple sources have become an obstacle, especially for users with basic or no programming skills. Here, we describe the difficulties related to handling SNP array data, focusing on file formats, SNP allele coding, and mapping. We also present SNPConvert suite, a multi-platform, open-source, and user-friendly set of tools to overcome these issues. This tool, which can be integrated with open-source and open-access tools already available, is a first step towards an integrated system to standardize and integrate any type of raw SNP array data. The tool is available at: https://github. com/nicolazzie/SNPConvert.git.

The effect of Vdr gene ablation on global gene expression in the mouse placenta.

The effects of vitamin D are mediated through the vitamin D receptor (VDR), a predominantly nuclear receptor, expressed in numerous tissues including the placenta. VDR and the retinoid X receptor (RXR) form a dimer complex which binds to genomic vitamin D responsive elements located primarily in promoter regions and recruit cell-specific transcription factor complexes which regulate the expression of numerous genes. To investigate the role of VDR on regulating placental gene expression, mice heterozygous (+/−) for an ablated Vdr allele (C57Bl6 strain B6.129S4-VDRtm1Mbd/J, Jackson Laboratory) were mated to generate Vdr+/+, Vdr+/− and Vdr −/− fetuses and placental samples were collected at day 18.5 of pregnancy. RNA was isolated from placental tissue with global gene expression measured using Affymetrix Mouse Gene 2.1 ST Arrays to assess the effects of VDR on global gene expression in the placenta. All raw array data are deposited in Gene Expression Omnibus (GEO) under accession GSE61583.

Abstract 347: Identification and Characterization of a miRNA Cohort Initiated Transitional Program That Controls Cell Cycle Arrest of the Perinatal Heart

During fetal and early perinatal development the myocardium undergoes a period of hyperplastic growth, which results in an exponential increase in the number of cardiomyocytes (CM) that will constitute the adult heart. Soon after birth, CMs proceed through a final round of cell division in the absence cytokinesis that results in binucleation of &gt;95% of adult CMs. Fetal heart genes are re-activated with the onset of pathological hypertrophic or dilated cardiomyopathies, yet there is no evidence of CM re-entry into the cell cycle. Despite the importance of this phenomenon, little is known about the molecular basis for the transition from hyperplastic to hypertrophic-based myocardial growth. Hypothesis: A perinatal heart gene program is necessary for the normal transition from a fetal heart gene program to an adult heart gene program. To identify the molecular mechanisms and pathways involved in CM differentiation during the perinatal transition, RNA was isolated from E18, and 1, 3, 5, 7, 10 and 35d old mouse hearts. CM gene expression and micro-RNA profiles (n=3 arrays/time point) were determined by oligonucleotide array analysis. The raw array data was normalized by Robust Multi-array analysis. Empirical Bayes estimation of gene-specific variances was performed between each of the time points in order to identify genes that are transiently and significantly changed at days 3 and 5 as compare to E18 and 10d post-birth. The analysis identified 2,799 genes (E18 v 5d) and 3,347 genes (5d v 10d) that were then clustered to determine significant pathway enrichment (p&lt;0.05) with Ingenuity Pathway Analysis. Our analysis confirmed previous observations of a down regulation of glucose oxidative metabolism (p=0.02) with an up-regulation of fatty acid metabolism (p=0.0001) between E18 and 5d post-birth. Also, 63 cell cycle genes are collectively down regulated (p=4.3x10-4) between 5d and 10d post-birth. We identified 131 genes that are transiently up regulated at 5d compared to E18 and 10d and this transition was proceeded by a specific cohort of miRNAs. The data generated from this study provide new insight into the molecular mechanisms by which CMs regulate and permanently exit from the cell cycle.

Transcriptomic analyses of Onecut1 and Onecut2 deficient retinas.

In this article, we further explore the data generated for the research article “Onecut1 and Onecut2 play critical roles in the development of the mouse retina”. To better understand the functionality of the Onecut family of transcription factors in retinogenesis, we investigated the retinal transcriptomes of developing and mature mice to identify genes with differential expression. This data article reports the full transcriptomes resulting from these experiments and provides tables detailing the differentially expressed genes between wildtype and Onecut1 or 2 deficient retinas. The raw array data of our transcriptomes as generated using Affymetrix microarrays are available on the NCBI Gene Expression Omnibus (GEO) browser (Reference number GSE57917 and GSE57918GSE57917GSE57918).