论文部分内容阅读
Hypoxia in tumours may confer resistance to conventional therapies and is associated with a poorer prognosis.MicroRNA expression alterations have been described in cancer and certain microRNAs have shown regulation by hypoxia.Here we integrate miRNA and mRNA gene expression in a breast cancer (BC) clinical dataset of 210 samples.The expression of 656 miRNAs and of 24,000 transcripts was measured using Illumina platforms in 210 early BCs with long term follow-up where Illumina mRNA expression microarrays were also available.We use previously derived gene expression signatures of BC sub-types and biological pathways to identify miRNAs associated with these subtypes and pathways, whilst correcting for the effect of all the others, and also to identify miRNAs that are prognostic,after accounting for the effect of these signatures and clinico-pathological factors.The incorporation of gene expression and clinico-pathological variability within the data-mining avoids artifacts occuring because of population heterogeneity.