Folate intake and colorectal cancer risk according to genetic subtypes defined by targeted tumor sequencing.

BACKGROUND: Folate is involved in multiple genetic, epigenetic, and metabolic processes, and inadequate folate intake has been associated with an increased risk of cancer. OBJECTIVE: We examined whether folate intake is differentially associated with colorectal cancer (CRC) risk according to somatic mutations in genes linked to CRC using targeted sequencing. DESIGN: Participants within two large CRC consortia with available information on dietary folate, supplemental folic acid, and total folate intake were included. Colorectal tumor samples from cases were sequenced for the presence of non-silent mutations in 105 genes and 6 signaling pathways (IGF2/PI3K, MMR, RTK/RAS, TGF-ß, WNT, TP53/ATM). Multinomial logistic regression models were run comparing mutated/non-mutated CRC cases to controls to compute multivariable-adjusted odds ratios (ORs) with 95% confidence intervals (CI). Heterogeneity of associations of mutated versus non-mutated CRC cases was tested in case-only analyses using logistic regression. Analyses were performed separately in hypermutated and non-hypermutated tumors, as they exhibit different clinical behaviors. RESULTS: We included 4,339 CRC cases (702 hypermutated tumors, 16.2%) and 11,767 controls. Total folate intake was inversely associated with CRC risk (OR=0.93, 95%CI=0.90-0.96). Among hypermutated tumors, 12 genes (AXIN2, B2M, BCOR, CHD1, DOCK3, FBLN2, MAP3K21, POLD1, RYR1, TET2, UTP20, ZNF521) showed nominal statistical significance (P<0.05) for heterogeneity by mutation status, but none remained significant after multiple testing correction. Among these genetic subtypes, the associations between folate variables and CRC were mostly inverse or towards the null, except for tumors mutated for DOCK3 (supplemental folic acid), CHD1 (total folate), and ZNF521 (dietary folate) that showed positive associations. We did not observe differential associations in analyses among non-hypermutated tumors, or according to the signaling pathways. CONCLUSIONS: Folate intake was not differentially associated with CRC risk according to mutations in the genes explored. The nominally significant differential mutation effects observed in a few genes warrants further investigation.

Association of gut microbial communities with plasma lipopolysaccharide-binding protein (LBP) in premenopausal women.

The mechanisms by which obesity increases cancer risk are unclear, but some lines of evidence suggest that gut microbial communities (GMC) may contribute to chronic inflammation in obese individuals through raised systemic levels of lipopolysaccharides (LPS). We evaluated associations of the GMC in stool with plasma LPS-binding protein (LBP, a measure of LPS) and C-reactive protein (CRP) concentrations in 110 premenopausal women in the United States. Diet was assessed using 3-day food records and GMCs were evaluated using pyrosequencing of the 16S rRNA gene. OTUs were identified at 97% sequence similarity. Taxonomic classification and functional genes were imputed from 16S rRNA genes, and alpha and beta diversity were assessed using the Shannon index and MRPP, respectively. Multivariable linear regression analysis was used to assess the relation between LBP, specific bacterial genera identified with indicator species analysis, and CRP. Dietary fat intake, particularly saturated fat, and CRP were positively associated with increased LBP. GMC beta diversity, but not alpha diversity, was statistically significantly different between groups using unweighted Unifrac. Several taxa, particularly those in the Clostridia class, were more prevalent in women with low LBP, while Bacteroides were more prevalent in those with high LBP. Genes associated with gram-negative cell wall material synthesis were also associated with LBP and CRP. In contrast, Phascolarctobacterium was associated with lower concentrations of LBP and CRP. We found distinct differences between tertiles of LBP regarding the diversity and composition of the microbiome, as well as differences in functional genes that potentially activate LBP.

The Model-Based Study of the Effectiveness of Reporting Lists of Small Feature Sets Using RNA-Seq Data.

Ranking feature sets for phenotype classification based on gene expression is a challenging issue in cancer bioinformatics. When the number of samples is small, all feature selection algorithms are known to be unreliable, producing significant error, and error estimators suffer from different degrees of imprecision. The problem is compounded by the fact that the accuracy of classification depends on the manner in which the phenomena are transformed into data by the measurement technology. Because next-generation sequencing technologies amount to a nonlinear transformation of the actual gene or RNA concentrations, they can potentially produce less discriminative data relative to the actual gene expression levels. In this study, we compare the performance of ranking feature sets derived from a model of RNA-Seq data with that of a multivariate normal model of gene concentrations using 3 measures: (1) ranking power, (2) length of extensions, and (3) Bayes features. This is the model-based study to examine the effectiveness of reporting lists of small feature sets using RNA-Seq data and the effects of different model parameters and error estimators. The results demonstrate that the general trends of the parameter effects on the ranking power of the underlying gene concentrations are preserved in the RNA-Seq data, whereas the power of finding a good feature set becomes weaker when gene concentrations are transformed by the sequencing machine.