Prediction Errors in Learning Drug Response from Gene Expression Data – Influence of Labeling,Sample Size,and Machine Learning Algorithm

Model-based prediction is dependent on many choices ranging from the sample collection and prediction endpoint to the choice of algorithm and its parameters. Here we studied the effects of such choices, exemplified by predicting sensitivity (as IC50) of cancer cell lines towards a variety of compounds. For this, we used three independent sample collections and applied several machine learning algorithms for predicting a variety of endpoints for drug response. We compared all possible models for combinations of sample collections, algorithm, drug, and labeling to an identically generated null model. The predictability of treatment effects varies among compounds, i.e. response could be predicted for some but not for all. The choice of sample collection plays a major role towards lowering the prediction error, as does sample size. However, we found that no algorithm was able to consistently outperform the other and there was no significant difference between regression and two- or three class predictors in this experimental setting. These results indicate that response-modeling projects should direct efforts mainly towards sample collection and data quality, rather than method adjustment.     

Genome-Wide Profiling of p63 DNA–Binding Sites Identifies an Element that Regulates Gene Expression during Limb Development in the 7q21 SHFM1 Locus

Heterozygous mutations in p63 are associated with split hand/foot malformations (SHFM), orofacial clefting, and ectodermal abnormalities. Elucidation of the p63 gene network that includes target genes and regulatory elements may reveal new genes for other malformation disorders. We performed genome-wide DNA–binding profiling by chromatin immunoprecipitation (ChIP), followed by deep sequencing (ChIP–seq) in primary human keratinocytes, and identified potential target genes and regulatory elements controlled by p63. We show that p63 binds to an enhancer element in the SHFM1 locus on chromosome 7q and that this element controls expression of DLX6 and possibly DLX5, both of which are important for limb development. A unique micro-deletion including this enhancer element, but not the DLX5/DLX6 genes, was identified in a patient with SHFM. Our study strongly indicates disruption of a non-coding cis-regulatory element located more than 250 kb from the DLX5/DLX6 genes as a novel disease mechanism in SHFM1. These data provide a proof-of-concept that the catalogue of p63 binding sites identified in this study may be of relevance to the studies of SHFM and other congenital malformations that resemble the p63-associated phenotypes.     

Histochemical Discrimination of Endogenous Mammalian β-galactosidase Activity from that Resulting from lac-Z Gene Expression

Minces of several organs from the transgenic mouse ROSA-gal 26 (ROSA-26), which robustly expresses bacterial lac-Z in most tissues, were exposed to 4-bromo, 5-chloro, 3-indoyl, -D-galactopyrosanide (X-gal) at pH ranging from 7.5 to 9.5 to determine the optimal pH for in situ demonstration of bacterial -galactosidase activity (neutral pH optimum) while minimizing detection of potentially confounding endogenous mammalian -galactosidase (acidic pH optimum). Similar studies were performed with organ minces from C57BL/6 mice, Sprague-Dawley rats, New Zealand white rabbits, and macaques to confirm the effect of pH on minimizing detection of endogenous mammalian -galactosidase. In all organs evaluated; heart, liver, spleen, kidney, brain, and skeletal muscle, endogenous -galactosidase activity was rarely detected following incubation at pH greater than 7.5. In contrast, bacterial -galactosidase activity in the ROSA-26 mice was strongly detected in organ minces following incubation at pH 8.0–9.0. These findings are similar to previous observations we have made in lung minces and confirm that a simple alteration of a commonly used histochemical technique for detecting in situ -galactosidase activity, raising the reaction buffer pH to weakly alkaline range, can reliably distinguish between endogenous activity and that resulting from exogenous bacterial gene expression.     

Cloning and Characterization of the nagA Gene that Encodes β-N-Acetylglucosaminidase from Aspergillus nidulans and Its Expression in Aspergillus oryzae

We isolated a β-N-acetylglucosaminidase encoding gene and its cDNA from the filamentous fungus Aspergillus nidulans, and designated it nagA. The nagA gene contained no intron and encoded a polypeptide of 603 amino acids with a putative 19-amino acid signal sequence. The deduced amino acid sequence was very similar to the sequence of Candida albicans Hex1 and Trichoderma harzianum Nag1. Yeast cells containing the nagA cDNA under the control of the GAL1 promoter expressed β-N-acetylglucosaminidase activity. The chromosomal nagA gene of A. nidulans was disrupted by replacement with the argB marker gene. The disruptant strains expressed low levels of β-N-acetylglucosaminidase activity and showed poor growth on a medium containing chitobiose as a carbon source. Aspergillus oryzae strain carrying the nagA gene under the control of the improved glaA promoter produced large amounts of β-N-acetylglucosaminidase in a wheat bran solid culture.     

Temperature-Shift Analysis of Bacteriophage φ29 Gene Expression in Bacillus amyloliquefaciens

Temperature shift-up experiments with conditional lethal mutants of Bacillus phage phi29 have allowed placement of early, middle, and late functions on the linear phi29 genetic map Most of the phi29 cistrons are late and are found at the ends of the map.     

Transformation and Functional Expression of the rFCA-RRM2 Gene in Rice

The primary aim of the present study was to investigate the overexpression of the rice (Oryza sativa L. subsp, japonica var. Zhonghua 11) flowering control gene (rFCA-RRM2) in monocotyledonous model rice. Constitutive expression of rFCA-RRM2 from the Actl-5 rice promoter caused late flowering in transgenic rice and increased grain weight that was more than 50% higher than that of control plants, which is the first demonstration of rFCA-RRM2 being able to increase rice production. Late flowering was accompanied by strong phenotype and some morphological modifications. These observations suggest that rFCA-RRM2 is a useful tool for phenotype improvement and yield enhancement in cereal crops.     

Sox9 Represses α-Sarcoglycan Gene Expression in Early Myogenic Differentiation

Alpha sarcoglycan (α-SG) is highly expressed in differentiated striated muscle, and its disruption causes limb-girdle muscular dystrophy. Accordingly, the myogenic master regulator MyoD finely modulates its expression. However, the mechanisms preventing α-SG gene expression at early stages of myogenic differentiation remain unknown. In this study, we uncovered Sox9, which was not previously known to directly bind muscle gene promoters, as a negative regulator of α-SG gene expression. Reporter gene and chromatin immunoprecipitation assays revealed three functional Sox-binding sites that mediate α-SG promoter activity repression during early myogenic differentiation. In addition, we show that Sox9-mediated inhibition of α-SG gene expression is independent of MyoD. Moreover, we provide evidence suggesting that Smad3 enhances the repressive activity of Sox9 over α-SG gene expression in a transforming growth factor-β-dependent manner. On the basis of these results, we propose that Sox9 and Smad3 are responsible for preventing precocious activation of α-SG gene expression during myogenic differentiation.     

Seasonal Variations in Clock‐Gene Expression in Atlantic Salmon (Salmo salar)

In homeothermic vertebrates inhabiting temperate latitudes, it is clear that the seasonal changes in daylength are decoded by the master circadian clock, which through secondary messengers (like pineal melatonin secretion) entrains rhythmic physiology to local conditions. In contrast, the entrainment and neuroendocrine regulation of rhythmic physiology in temperate teleosts is not as clear, primarily due to the lack of understanding of the clock gene system in these species. In this study, we analyzed the diel expression of the clock‐genes in brains of Atlantic salmon, a species that is both highly photoperiodic and displays robust clock‐controlled behavior. Atlantic salmon parr were acclimated to either long‐day (LD) or short‐day (SD) photoperiods for one month and thereafter sampled at 4 h intervals over a 24 h cycle. Clock, Bmal1, Per2, and Cry2 were all actively expressed in salmon brain homogenates and, with the exception of Per2, all displayed rhythmic expression under SD photoperiods that parallels that reported in zebrafish. Interestingly, daylength significantly altered the mRNA expression of all clock genes studied, with Clock, Bmal1, and Per2 all becoming arrhythmic under the LD compared to SD photoperiod, while Cry2 expression was phase delayed under LD. It is thus proposed that the clock‐gene system is actively expressed in Atlantic salmon, and, furthermore, as has been reported in homeothermic vertebrates, it appears that clock expression is daylength‐dependent.