The transfectants were subjected in immunostaining using antiH3K27me3 and anti-Flag antibodies

creased trabecular separation and higher turnover compared to that observed in dabigatran-treated or control rats. These findings suggest that dabigatran has a better bone safety profile than warfarin. Since warfarin treatment affects bone by reducing trabecular size and structure, increasing turnover and reducing mineralization, these differences could translate into a lower incidence of fractures in dabigatran treated patients. ~~ With the recent advances in next-generation sequencing technologies, the prospects of personalized healthcare look brighter than ever. The use of genomics to guide clinical care is perhaps most widespread in cancer. Many pioneer studies have shown how one can use signatures of gene expression to predict clinical outcomes for individual patients. More recently two large collections of matched drug screens and genomics profiles of cancer cell lines have been published. These data have been used to build predictive models of drug response by associating genomic features with drug sensitivity in cancer cell lines. Additionally, connecting drug sensitivity to specific genomic features can help shed light on the mechanisms of drug action and elucidate the underlying reasons for resistance to the treatment. Thus, these data offer the opportunity to develop methods that can be used for personalized treatment. 1 / 22 Context MedChemExpress R-7128 Sensitive Modeling of Cancer Drug Sensitivity A key challenge in associating genetic characteristics to drug sensitivity is the role of context in biological systems. For example, regulation of gene expression has been shown to have patterns specific to tissues and cell-types. In tumorigenesis, diverse patterns of mutation, gene expression, and epigenetic regulation have also been observed in cancer-specific or tissuespecific manner. This context dependency plays an important role in the efficacy of treatment. For example, PLX4732, a RAF inhibitor targeting oncogenic BRAFV600E, is a potent treatment for melanoma patients with the mutation. However, colon cancer patients with the same mutation do not respond to PLX4732. It is therefore important to take into account the context created by cancer types when analyzing the genomics of drug sensitivity. It is no surprise that predictive models built using only melanoma data give better PubMed ID:http://www.ncbi.nlm.nih.gov/pubmed/19752305 title=View Abstract(s)”>PubMed ID:http://www.ncbi.nlm.nih.gov/pubmed/19752319 prediction for melanoma samples than those built using data of mixed cancer types. This argues that we should focus on one cancer type when building models for drug sensitivity. While such strategy allows us to avoid confounding influence of context, it constrains us to a small number of samples. Due to sample size, current datasets lack the statistical power to build separate models for each cancer. We utilize commonality between cancer types and drugs to overcome the paucity of data. We propose CHER, an algorithm that builds predictive models by selecting genomic features and deciding which ones are shared or not between cancer types, tissues and drugs. CHER is empowered by two assumptions. First, CHER assumes similar cancer types may have similar mechanisms underlying drug sensitivity. For example, basal-like breast cancer and ovarian cancer share many molecular signatures; therefore, these two cancers are likely to share similar predictive genomic features for drug sensitivity. Second, CHER assumes that if two drugs induce similar responses, their predictive models are likely similar. These assumptions allow CHER to boost its power to uncover biomarkers predictive o