Chen, Jun

Abstract Background: The need to develop valid methods for sampling and analyzing fecal specimens for microbiome studies is increasingly important, especially for large population studies. Methods: Some of the most important attributes of any sampling method are reproducibility, stability, and accuracy. We compared seven fecal sampling methods [no additive, RNAlater, 70% ethanol, EDTA, dry swab, and pre/post development fecal occult blood test (FOBT)] using 16S rRNA microbiome profiling in two laboratories. We evaluated nine commonly used microbiome metrics: abundance of three phyla, two alpha-diversities, and four beta-diversities. We determined the technical reproducibility, stability at ambient temperature, and accuracy. Results: Although microbiome profiles showed systematic biases according to sample method and time at ambient temperature, the highest source of variation was between individuals. All collection methods showed high reproducibility. FOBT and RNAlater resulted in the highest stability without freezing for 4 days. In comparison with no-additive samples, swab, FOBT, and 70% ethanol exhibited the greatest accuracy when immediately frozen. Conclusions: Overall, optimal stability and reproducibility were achieved using FOBT, making this a reasonable sample collection method for 16S analysis. Impact: Having standardized method of collecting and storing stable fecal samples will allow future investigations into the role of gut microbiota in chronic disease etiology in large population studies. Cancer Epidemiol Biomarkers Prev; 25(2); 407–16. ©2015 AACR.

Classic multinomial logit model, commonly used in multiclass regression problem, is restricted to few predictors and does not take into account the relationship among variables. It has limited use for genomic data, where the number of genomic features far exceeds the sample size. Genomic features such as gene expressions are usually related by an underlying biological network. Efficient use of the network information is important to improve classification performance as well as the biological interpretability. We proposed a multinomial logit model that is capable of addressing both the high dimensionality of predictors and the underlying network information. Group lasso was used to induce model sparsity, and a network-constraint was imposed to induce the smoothness of the coefficients with respect to the underlying network structure. To deal with the non-smoothness of the objective function in optimization, we developed a proximal gradient algorithm for efficient computation. The proposed model was compared to models with no prior structure information in both simulations and a problem of cancer subtype prediction with real TCGA (the cancer genome atlas) gene expression data. The network-constrained mode outperformed the traditional ones in both cases.