Computational protocol: Transcriptomic profiling of Melon necrotic spot virus-infected melon plants revealed virus strain and plant cultivar-specific alterations

Similar protocols

Protocol publication

[…] The hybridization data were obtained from our experiment and from public repositories. Our experiment data provided by NimbleGen, were grouped for normalization into two groups: cotyledon and leaf. As the CMV data came from an older platform, each hybridization data group (CMV, WMV and MNSV-α5 at 3dpi) was independently normalized for later analysis and comparison. Each group of data were normalized and transformed to a log2 scale using the RMA (Robust Multi-array Average) algorithm found in the oligo package [] of Bioconductor (http://www.bioconductor.org). For the cotyledon time-course experiment, the maSigPro package [] was used to identify differentially-expresed genes. This program uses a two-regression step strategy. In the first step, a general regression model is defined. Then, the defined model is adjusted to the data through least squares, and the genes that significantly differ from this regression model are identified by correcting with a specific false discovery rate (FDR) of 1 % (Q = 0.01). In the second step, a stepwise regression is employed, and a probability (p) is calculated for each variable, showing the probability that causes the deviation. After the analysis, a list of the differentially expressed genes is obtained according to each variable (“TIME”, “TIME x Virus” and “Virus vs. Control”). We discard the differentially expressed genes associated with only the variable “TIME” in order to select those genes that were deregulated with time and associated with the virus in each cultivar. A maSigPro analysis was conducted for each cultivar. For the single stage experiment, the identification of differentially expressed genes was done through the SAM (Significant analysis of microarrays) module [] found in the Multi Experimental Viewer (MeV, v. 4.9.0) program [], using a FDR = 0. Genes with a fold change smaller than 2 (FC ≤2, cut-off of log2 ≤ 1) were filtered out.Samples were grouped with the PCA module from MeV []. Clustering of the samples was done with Euclidean distance by hierachical clustering [] and the bootstrap was done by Support trees [] (bootstrap 100 replicates). Genes were clustered by their expression pattern by using the k-means clustering method [] and Pearson’s correlation for the calculation of distances. Lastly, the functional analysis was done with the Blast2GO program [], extracting the over- or under-represented GO terms among the differentially-expressed genes from each condition by the application of Fisher’s test (p-value <0.05). [...] The melon microarray was validated in previous works [–] and we undertook further verification by comparing microarray and RT-qPCR expression patterns of a pathogen response protein, a calmodulin-binding protein, a lipoxigenase and a glucosyl transferase transcript. The same RNA samples from cotyledon hybridized to the microarray were used for this purpose. Data from RT-qPCR were transformed to a log2 scale to make the data comparable with microarray results. A strong positive correlation was found between the two sets of values (R2 = 0.86; correlation coefficient of 0.93) (Additional file ), confirming previous results [, ].For real time quantitative PCR, the first strand cDNA was synthesized using 1.5 μg of total RNA, following the directions of the reverse transcriptase manufacturer (Roche) with an oligo-dT(16) as reverse primer. As MNSV and CMV do not have a poly(A) tail, reverse primers for the respective viruses (CE-948, 5′-CCCACTATCATCACGATCTTTAC-3′, and CE-169, 5′-CCGCTTACGATTCCCAACTGT-3′) were added for transcription of the viral RNAs. The qPCR for the quantification of messenger RNA, as well as viral accumulation was performed on an AB7500 System (Applied Biosystems), using SYBR Green PCR Master Mix (Applied Biosystems) as the detector and ROX as the passive reference. All the reactions (final volume of 20 μl) contained 10 μl Master Mix, 0.15 μl of each primer (100 mM) and 60 ng of cDNA. Each reaction was done in triplicate, along with controls without DNA (NTC), using a two-step amplification protocol and adding a melting curve. The analysis of the melting curves and the NTC were done in order to ensure the specific amplification of the product and the absence of dimerization of the primers. The primers used for amplification of the target and reference genes are listed in Additional file .For calculating the relative quantification of each transcript, we used the 2ΔΔct method. The relative expression levels were determined through the normalization of the samples with mRNA from cyclophilin (cCL3169Contig1) as an internal control and relating it to the expression values of the healthy controls. The analysis was carried out with the SDS-7500 software and exported to a spread sheet for further calculations. The specific primer pairs were designed with Primer Express software v3.0 (Applied Biosystems). The efficiency of each primer pair was calculated through the equation: Efficiency (%) = (10[-1/slope] - 1) x 100 (Guide to performing relative quantitation of gene expression using real-time quantitative PCR, Applied Biosystems). […]

Pipeline specifications

Software tools Oligo package, maSigPro, Blast2GO, Primer Express
Applications Gene expression microarray analysis, qPCR
Organisms Cucumis melo, Cucumber mosaic virus, Watermelon mosaic virus
Diseases Infection, HIV Infections