Selection bias in gene extraction on the basis of microarray gene-expression data

In the context of cancer diagnosis and treatment, we consider the problem of constructing an accurate prediction rule on the basis of a relatively small number of tumor tissue samples of known type containing the expression data on very many (possibly thousands) genes. Recently, results have been presented in the literature suggesting that it is possible to construct a prediction rule from only a few genes such that it has a negligible prediction error rate. However, in these results the test error or the leave-one-out cross-validated error is calculated without allowance for the selection bias. There is no allowance because the rule is either tested on tissue samples that were used in the first instance to select the genes being used in the rule or because the cross-validation of the rule is not external to the selection process; that is, gene selection is not performed in training the rule at each stage of the cross-validation process. We describe how in practice the selection bias can be assessed and corrected for by either performing a cross-validation or applying the bootstrap external to the selection process. We recommend using 10-fold rather than leave-one-out cross-validation, and concerning the bootstrap, we suggest using the so-called. 632+ bootstrap error estimate designed to handle overfitted prediction rules. Using two published data sets, we demonstrate that when correction is made for the selection bias, the cross-validated error is no longer zero for a subset of only a few genes.

Suicide attempts by psychiatric patients in acute inpatient, long-stay inpatient and community care

Background: This study examined rates of and risk factors associated with suicide attempts by psychiatric patients under active care. It was especially focussed on the relative rates across three standard treatment settings: acute inpatient care, long-stay inpatient care and community-based carl. Methods: A total of 12,229 patients in 13,632 episodes of care were rated on the Health of the Nation Outcome Scales (HoNOS) Item 2. For the purposes of the current investigation, a score of 4 was deemed to indicate a suicide attempt. Results: Incidence densities per 1000 episode days were 5.4 (95% CI = 4.8-6.1) for patients under care in acute inpatient settings, 0.6 (95% CI = 0.5-0.8) for patients under care in long-stay inpatient settings, and 0.5 (95% CI = 0.5-0.6) for patients under carl in community-based arrangements. Predictors varied by treatment setting. Risk was elevated for personality disorders across all settings: 22.7 attempts per 1000 episode days (95% CI = 17.2-30.0) in acute inpatient care; 2.1 (95% CI = 1.0-4.5) in long-stay inpatient care; and 2.3 (95% CI = 1.7-3.0) in community-based care. This effect remained after adjustment for demographics. Conclusion: Rates of suicide attempts among psychiatric patients are a major issue facing contemporary mental health care systems, and risk factors vary across different treatment settings.

The first non-turnover voltammetric response from a molybdenum enzyme: direct electroscopy of dimethylsulfoxide reductase from Rhodobacter capsulatus

The first direct voltammetric response from a molybdenum enzyme under non-turnover conditions is reported. Cyclic voltammetry of dimethylsulfoxide reductase from Rhodobacter capsulatus reveals a reversible Mo-VI/V response at + 161 mV followed by a reversible Mo-V/IV response at -102 mV versus NHE at pH 8. The higher potential couple exhibits a pH dependence consistent with protonation upon reduction to the Mo-V state and we have determined the pK(a) for this semi-reduced species to be 9.0. The lower potential couple is pH independent within the range 5 < pH < 10. The optical spectrum of the Mo chromophore has been investigated with spectroelectrochemistry. At high potential, in its resting state, the enzyme exhibits a spectrum characteristic of the Mo-VI form. This changes significantly following bulk electrolysis (-400 mV versus NHE) at an optically transparent, indium-doped tin oxide working electrode, where a single visible electronic maximum at 632 nm is observed, which is comparable with spectra reported previously for the dithionite-reduced enzyme. This two-electron process is chemically reversible by reoxidizing the enzyme at the electrode in the absence of mediators or promoters. The activity of the enzyme has been established by observation of a catalytic current in the presence of DMSO at pH 8, where a sigmoidal (steady state) voltammogram is seen. Electronic supplementary material to this paper (Fig. S 1) can be obtained by using the Springer Link server located at http://dx.doi.org/10.1007/s00775-002-0374-y.