thaliana (TAIR9, Swarbreck et al , 2008) and O sativa (Rice Geno

thaliana (TAIR9, Swarbreck et al., 2008) and O. sativa (Rice Genome Annotation Project v6.1, Ouyang

et al., 2007) via BLASTX (for a complete workflow see Fig. S4). Gene-expression profiles were analyzed by multivariate analysis to E7080 research buy identify similarities and differences of the entire transcriptomic response between species and treatment conditions. Transcription profiles of the eight libraries were normalized for library size and composition of expressed transcripts (Robinson and Oshlack, 2010). Groupings of expression profiles based on the biological coefficient of variation between library pairs were identified with multidimensional-scaling (MDS) using the R package “edgeR” v2.5.1 (Robinson et al., 2010). Identified groupings were tested by ANOSIM analysis (analysis of similarity, tests distances within vs. between groups) implemented in the R package “vegan” v2.0–3 (Oksanen et al., 2012). Multivariate analysis and subsequent expression analysis along with plotting functions were performed in R (R Development Core Team, 2008). Differential expression analysis was performed with the R package “edgeR”, which employs an overdispersed Poisson model (negative binomial) to account for technical and biological variability,

with the generalized linear model (GLM) functionality for multifactor experiments (Robinson et al., 2010 and McCarthy et al., 2012). Differentially expressed genes were determined for three data sets: 1) eight libraries including samples of both species, 2) four libraries this website of Z. marina and 3) four libraries of N. noltii. In all three data sets, the expression profiles were normalized for library size and composition of expressed transcripts ( Robinson and Oshlack,

2010). For the data set including both species (data set 1), the single factor species was fitted to the GLM to test for differential expression between both species consistent across treatments. In this case, all four libraries per species from the two different populations and treatments were used as biological replicates on the species level. For Z. marina alone L-NAME HCl (data set 2) the data were analyzed with GLM including the factors treatment and population (the factor population was suggested by the grouping of expression profiles; Fig. 1). Differential expression, with respect to heat treatment, was tested, while adjusting for the remaining factor. For N. noltii alone (data set 3) the factor “group identity” with three factor levels identified by MDS ( Fig. 1) was fitted to the GLM. Genes displaying differential expression between heat and control treatment in the northern population (two of the three groups, Fig. 1) were identified. In all three data sets, the biological replication as defined by the design of the respective GLM was used to calculate the tagwise dispersion, the overdispersion value in the negative binomial model ( Robinson et al., 2010 and McCarthy et al., 2012).

Leave a Reply

Your email address will not be published. Required fields are marked *

*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>