Search

As new technologies permit the generation of hitherto unprecedented volumes of data (e.g. genome-wide association study data), researchers struggle to keep up with the added complexity and time commitment required for its analysis. For this reason, model selection commonly relies on machine learning and data-reduction techniques, which tend to afford models with obscure interpretations. Even in cases with straightforward explanatory variables, the so-called ‘best’ model produced by a given model-selection technique may fail to capture information of vital importance to the domain-specific questions at hand. Herein we propose a new concept for model selection, feasibility, for use in identifying multiple models that are in some sense optimal and may unite to provide a wider range of information relevant to the topic of interest, including (but not limited to) interaction terms. We further provide an R package and associated Shiny Applications for use in identifying or validating feasible models, the performance of which we demonstrate on both simulated and real-life data.

The flood basalts and gold-bearing basal sediments of the 2775-2629 Ma Fortescue Group unconformably overlie the Mesoarchean West Pilbara Superterrane to the south of Karratha, Western Australia. F...

From birth to adulthood, the gut microbiota matures from a simple community dominated by a few major bacterial groups into a highly diverse ecosystem that provides both benefits and challenges to the host. Currently there is great interest in identifying environmental and host factors that shape the development of our gut microbiota. Breast milk is a rich source of maternal antibodies, which provide the first source of adaptive immunity in the newborn's intestinal tract. In this addendum, we summarize our recent data demonstrating that maternal antibodies in breast milk promote long-term intestinal homeostasis in suckling mice by regulating the gut microbiota and host gene expression. We also discuss important unanswered questions, future directions for research in this field, and implications for human health and disease.

MicroRNAs (miRNAs) guide Argonaute (AGO)-containing microribonucleoprotein (miRNP) complexes to target mRNAs. It has been assumed that miRNAs behave similarly to each other with regard to mRNA target recognition. The usual assumptions, which are based on prior studies, are that miRNAs target preferentially sequences in the 3′UTR of mRNAs, guided by the 5′ “seed” portion of the miRNAs. Here we isolated AGO- and miRNA-containing miRNPs from human H4 tumor cells by co-immunoprecipitation (co-IP) with anti-AGO antibody. Cells were transfected with miR-107, miR-124, miR-128, miR-320, or a negative control miRNA. Co-IPed RNAs were subjected to downstream high-density Affymetrix Human Gene 1.0 ST microarray analyses using an assay we validated previously—a “RIP-Chip” experimental design. RIP-Chip data provided a list of mRNAs recruited into the AGO-miRNP in correlation to each miRNA. These experimentally identified miRNA targets were analyzed for complementary six nucleotide “seed” sequences within the transfected miRNAs. We found that miR-124 targets tended to have sequences in the 3′UTR that would be recognized by the 5′ seed of miR-124, as described in previous studies. By contrast, miR-107 targets tended to have ‘seed’ sequences in the mRNA open reading frame, but not the 3′ UTR. Further, mRNA targets of miR-128 and miR-320 are less enriched for 6-mer seed sequences in comparison to miR-107 and miR-124. In sum, our data support the importance of the 5′ seed in determining binding characteristics for some miRNAs; however, the “binding rules” are complex, and individual miRNAs can have distinct sequence determinants that lead to mRNA targeting.

Marginal intake of dietary zinc can be associated with increased risk of cardiovascular diseases. In the current study we hypothesized that vascular dysfunction and associated inflammatory events are activated during a zinc deficient state.We tested this hypothesis using both vascular endothelial cells and mice lacking the functional LDL-receptor gene.Zinc deficiency increased oxidative stress and NF-kappaB DNA binding activity, and induced COX-2 and E-selectin gene expression, as well as monocyte adhesion in cultured endothelial cells. The NF-kappaB inhibitor CAPE significantly reduced the zinc deficiency-induced COX-2 expression, suggesting regulation through NF-kappaB signaling. PPAR can inhibit NF-kappaB signaling, and our previous data have shown that PPAR transactivation activity requires adequate zinc. Zinc deficiency down-regulated PPARalpha expression in cultured endothelial cells. Furthermore, the PPARgamma agonist rosiglitazone was unable to inhibit the adhesion of monocytes to endothelial cells during zinc deficiency, an event which could be reversed by zinc supplementation. Our in vivo data support the importance of PPAR dysregulation during zinc deficiency. For example, rosiglitazone induced inflammatory genes (e.g., MCP-1) only during zinc deficiency, and adequate zinc was required for rosiglitazone to down-regulate pro-inflammatory markers such as iNOS. In addition, rosiglitazone increased IkappaBalpha protein expression only in zinc adequate mice. Finally, plasma data from LDL-R-deficient mice suggest an overall pro-inflammatory environment during zinc deficiency and support the concept that zinc is required for proper anti-inflammatory or protective functions of PPAR.These studies suggest that zinc nutrition can markedly modulate mechanisms of the pathology of inflammatory diseases such as atherosclerosis.

This article proposes and studies the performance in theory and practice of the least trimmed differences (LTD) linear regression estimator. The estimator minimizes the sum of the smallest quartile of the squares of the differences in each pair of residuals. We obtain the breakdown point, maxbias curve, and large-sample properties of a class of estimators including the LTD as special case. The LTD estimator has a 50% breakdown point and Gaussian efficiency of 66%—substantially higher than other common high-breakdown estimators such as least median of squares and least trimmed squares. The LTD estimator is difficult to compute, but can be performed using a “feasible solution” algorithm. Half-sample jackknifing is effective in producing standard errors. In simulations we find the LTD to be more stable than other high-breakdown estimators. In an example, the LTD still shows instability like other high-breakdown estimators when there are small changes in the data.

One reason that robust methods are often not used in practice is the lack of affordable easy to use software. This paper addresses this issue by providing instructions, examples, and code for using Fortran and LISPSTAT to compute MM-estimates and other robust regression estimates. LISPSTAT is a graphical computing environment available for Unix-based, Dos and Macintosh computers at no charge to the user.

In the past decade there has been a substantial increase in the number of introductory statistics courses taught at the undergraduate level. Many have argued successfully for the extensive use of writing in such courses in an attempt to highlight the interdisciplinary role of statistics and acknowledge that a good statistician must also be good at summarizing his or her analyses to nonstatisticians. This point was made by Radke-Sharpe, who went on to add that incorporating writing demands time, energy, and creativity, but that it is usually well worth the effort. This article discusses the efforts made by the authors to include writing in their courses, and some of the techniques that made the writing process painless and productive for both students and faculty.

Many statistical techniques require that computations be done on all subsets of size r in a data set of size n. Typically, this is done lexographically, i.e., with nested for-loops. If an exchange one point update formula is available, then it is used on the inner loop. In this paper we discuss a method of counting through all subsets of size r in a data set of size n by changing only one element between successive subsets. Such methods have been studied in the applied mathematics literature but are mostly unknown to statisticians. The advantage of such methods is that an update formula can be used at every step, thus potentially saving computation time. The method used to compute the next subset in the list requires some computation time, and thus the new method will only be faster if the update formula is sufficiently faster than doing the computation from scratch.

Recent research by Sakata and White (1995) presents the consistency and asymptotic normality of S-estimators in nonlinear regression. It is well known from research in linear regression that it is important to use a consistent high breakdown estimator as an initial estimate when computing an S-estimate. This paper presents the proof of the weak consistency of the least median of squares estimator in a nonlinear regression setting, thus suggesting that it is a reasonable choice for the starting value for computing S-estimates in nonlinear regression.

The breakdown point is considered an important measure of the robustness of a linear regression estimator. This article addresses the concept of breakdown in nonlinear regression. Because it is not invariant to nonlinear reparameterization, the usual definition of the breakdown point in linear regression is inadequate for nonlinear regression. The original definition of breakdown due to Hampel is more suitable for nonlinear problems but may indicate breakdown when the fitted values change very little. We introduce breakdown functions, which measure breakdown of the fitted values. Using the breakdown functions, we introduce a new definition of the breakdown point. For the linear regression model, our definition of the breakdown point coincides with the usual definition for linear regression as well as with Hampel's definition. For most nonlinear regression functions, we show that the breakdown point of the least squares estimator is 1/n. We prove that for a large class of unbounded regression func...

(2002). Robust Diagnostic Regression Analysis. Journal of the American Statistical Association: Vol. 97, No. 457, pp. 365-366.