truncatula line 2HA and Jemalong using the Qiagen RNeasy plant mini protein inhibitors kit. Total RNA was quantified using a NanoDrop ND 1000 Spectrophotometer. RNA with an absorbance A260 A280ratio 2. 0 was quality tested using the Agilent 2100 Bioanalyzer. Preparation of cRNA, hybridisation, and scanning of the Test3 arrays and Medi cago GeneChip were performed according to the manu facturers protocol. Briefly, double stranded cDNA was synthesised from 5 to 8g of each RNA sample via oligo T7 24 primer mediated reverse transcription. Biotin labelled cRNA was generated using the Enzo BioArray kit, purified using RNeasy spin columns, and then quantified by spectrophotometer. Inhibitors,Modulators,Libraries Inhibitors,Modulators,Libraries Fifteen to 20g of each biotin labelled fragmented cRNA sample was used to prepare 300L of hybridisation mixture.

Aliquots of each sample were hybridised onto Test3 arrays to check the quality of the samples prior to hybridisation onto the Medicago genome arrays. The arrays were Inhibitors,Modulators,Libraries washed with optimised wash protocols, stained with strepdavidin phycoerythrin followed by antibody ampli fication, and scanned with the Agilent GeneArray Scanner. Data pre processing Raw Affymetrix data were normalised with the GCRMA algo rithm including quantile normalisation and variance stabilisation, using the Affymetrix package of the bioconductor software. The normalised aver age of the replicates Inhibitors,Modulators,Libraries was then log transformed in base 2 to reduce the proportional relationship between random error and signal intensity. Differentially expressed probe sets were identified by evaluating the log2 ratio between the two conditions associated to a standard t test, adjusted for multiple testing by the False Discovery Rate approach.

All probe Inhibitors,Modulators,Libraries sets that differed more than to two fold with a t test P value 0. 05 were consid ered to be differentially expressed. The Significance Anal ysis R115777 of Microarrays two class unpaired analysis was also performed in order to identify a more exten sive list of differentially expressed genes, with the measure significant fold change set at 2. 0 and a false discovery rate 8. 4%. The expected proportion of significantly different features was set to 0. 95. addition, probe sets of the Affymetrix Medicago Genome Array were assigned to gene families described in the TAIR database and to transcription factor families pro vided by the Database of Arabidopsis Transcription Fac tors based on their sequence similarity with Arabidopsis thaliana proteins. Blastx was used to find the best match for the sequences repre senting each probe set. The differentially expressed sets of sequences were compared to the composition of each gene family to iden tify if a certain category was statistically over represented.