Transforming RNA-Seq Data to Improve the Performance of Prognostic Gene Signatures

Gene expression measurements have successfully been used for building prognostic signatures, i.e for identifying a short list of important genes that can predict patient outcome. Mostly microarray measurements have been considered, and there is little advice available for building multivariable risk prediction models from RNA-Seq data. We specifically consider penalized regression techniques, such as the lasso and componentwise boosting, which can simultaneously consider all measurements and provide both, multivariable regression models for prediction and automated variable selection. However, they might be affected by the typical skewness, mean-variance-dependency or extreme values of RNA-Seq covariates and therefore could benefit from transformations of the latter. In an analytical part, we highlight preferential selection of covariates with large variances, which is problematic due to the mean-variance dependency of RNA-Seq data. In a simulation study, we compare different transformations of RNA-Seq data for potentially improving detection of important genes. Specifically, we consider standardization, the log transformation, a variance-stabilizing transformation, the Box-Cox transformation, and rank-based transformations. In addition, the prediction performance for real data from patients with kidney cancer and acute myeloid leukemia is considered. We show that signature size, identification performance, and prediction performance critically depend on the choice of a suitable transformation. Rank-based transformations perform well in all scenarios and can even outperform complex variance-stabilizing approaches. Generally, the results illustrate that the distribution and potential transformations of RNA-Seq data need to be considered as a critical step when building risk prediction models by penalized regression techniques.     

Effects of double ligation of Stensen's duct on the rabbit parotid gland

Salivary gland duct ligation is an alternative to gland excision for treating sialorrhea or reducing salivary gland size prior to tumor excision. Duct ligation also is used as an approach to study salivary gland aging, regeneration, radiotherapy, sialolithiasis and sialadenitis. Reports conflict about the contribution of each salivary cell population to gland size reduction after ductal ligation. Certain cell populations, especially acini, reportedly undergo atrophy, apoptosis and proliferation during reduction of gland size. Acini also have been reported to de-differentiate into ducts. These contradictory results have been attributed to different animal or salivary gland models, or to methods of ligation. We report here a bilateral double ligature technique for rabbit parotid glands with histologic observations at 1, 7, 14, 30, 60 days after ligation. A large battery of special stains and immunohistochemical procedures was employed to define the cell populations. Four stages with overlapping features were observed that led to progressive shutdown of gland activities: 1) marked atrophy of the acinar cells occurred by 14 days, 2) response to and removal of the secretory material trapped in the acinar and ductal lumens mainly between 30 and 60 days, 3) reduction in the number of parenchymal (mostly acinar) cells by apoptosis that occurred mainly between 14–30 days, and 4) maintenance of steady-state at 60 days with a low rate of fluid, protein, and glycoprotein secretion, which greatly decreased the number of leukocytes engaged in the removal of the luminal contents. The main post- ligation characteristics were dilation of ductal and acinar lumens, massive transient infiltration of mostly heterophils (rabbit polymorphonuclear leukocytes), acinar atrophy, and apoptosis of both acinar and ductal cells. Proliferation was uncommon except in the larger ducts. By 30 days, the distribution of myoepithelial cells had spread from exclusively investing the intercalated ducts pre-ligation to surrounding a majority of the residual duct-like structures, many of which clearly were atrophic acini. Thus, both atrophy and apoptosis made major contributions to the post-ligation reduction in gland size. Structures also occurred with both ductal and acinar markers that suggested acini differentiating into ducts. Overall, the reaction to duct ligation proceeded at a considerably slower pace in the rabbit parotid glands than has been reported for the salivary glands of the rat.     

Unterbuchungen über den Sitz des Géhörsinnes bei der Elritze

Ohne Zusammenfassung     

Comment on "Female reproductive development: a hazards model analysis," by Andrew Foster, Jane Menken, Aladdin Chowdhury and James Trussell

This article summarizes the differences between 2 similar studies. Endocrinological studies of the time to the 1st ovulatory cycle in early and late maturing girls in Finland (Apter and Vihko, 1983) are contrary to the Bangladeshi results reported by Foster in 1986. Apter found that the frequency of ovulation depended significantly on both the time since menarche and the age at menarche. Early menarche was associated with early onset of ovulatory cycles. It would be interesting to know why the Bangladesh data differ. They are puzzling, and their use in fertility models is problematic. This article also makes 2 corrections in the use of the fatness index in the Foster paper.     

Lysine-insensitive aspartate kinase in two threonine-overproducing mutants of maize

Aspartate kinase (AK; EC 2.7.2.A) catalyzes the first reaction in the biosynthesis pathway for aspartate-derived amino acids in plants. Aspartate kinase was purified from wildtype and two maize (Zea mays L.) genotypes carrying unlinked dominant mutations,Ask LT19 andAsk2 -LT20, that conferred overproduction of threonine, lysine, methionine and isoleucine. The objective of this investigation was to characterize the AKs from mutant and wildtype plants to determine their role in regulating the synthesis of aspartate-derived amino acids in maize. Kernels of the homozygousAsk2 mutant exhibited 174-, 10-, 13- and 2-fold increases in, in this sequence, free threonine, lysine, methionine and isoleucine, compared to wildtype. In wildtype maize, AK was allosterically feedback-inhibited by lysine with 10 μMl-lysine required for 50% inhibition. In contrast, AK purified from the isogenic heterozygousAsk and homozygousAsk2 mutants required 25 and 760 μM lysine for 50% inhibition, respectively, indicating thatAsk andAsk2 were separate structural loci for lysine-regulated AK subunits in maize. Further characterization of purified AK from the homozygous mutantAsk2 line indicated altered substrate and lysine inhibition kinetics. The apparent Hill coefficient was 0.7 for the mutantAsk2 AK compared with 1.6 for the wildtype enzyme, indicating that the mutant allele conferred the loss of a lysinebinding site to the mutant AK. Lysine appeared to be a linear noncompetitive inhibitor ofAsk2 AK with respect to MgATP and an uncompetitive inhibitor with respect to aspartate compared to S-parabolic, I parabolic noncompetitive inhibition of wildtype AK. Reduced lysine sensitivity of theAsk2 gene product appeared to reduce the lysine inhibition of all of the AK activity detected in homozygousAsk2 plants, indicating that maize AK is a heteromeric enzyme consisting of the two lysine-sensitive polypeptides derived from theAsk andAsk2 structural genes. Scientific paper No. 17419, Minnesota Agricultural Experiment Station projects No. 0302-4813-56 and No. 0302-4818-32 This research was supported in part by the U.S. Depatment of Agriculture Competitive Research Grants Office grant 86-CRCR-1-2019. The authors are grateful to Charles Grissom for providing the computer programs in an IBM-PC format.     

Tissue culture isolation of a second mutant locus for increased threonine accumulation in maize

Summary Regenerable maize (Zea mays L.) tissue cultures were selected for ability to grow in the presence of inhibitory (1.0–1.5 mM) concentrations of L-lysine plus L-threonine. Testcross kernels from one regenerated plant (LT20) segregated for wild-type and high free threonine concentration in a 11 ratio consistent with a single dominant gene for high free threonine. Free threonine concentrations (nmol/mg dry weight) increased an average of 29-fold in bulked F₂ kernel samples from heterozygous mutant plants, and the total (free plus protein-bound) threonine concentration increased 68%. Increases in protein-bound methionine, lysine and glycine concentrations were also noted, suggesting a possible effect of the mutation on protein concentration and composition. Allelism tests with a previously selected mutant line, Ltr ^*19, showed that two unlinked, codominant genes conditioned the high free threonine phenotype. Based on a separate study of aspartate kinase feedback inhibition characteristics in the two mutant lines, we propose that the mutant alleles [gene and allele designations are according to guidelines for maize genetic nomenclature (Burnham et al. 1975)] be designated Ask-LT19 and Ask2-LT20 for the Ltr ^*19 and LT20 mutants, respectively.