Ida Scheel

The influence of missing value imputation on detection of differentially expressed genes from microarray data

MOTIVATION Missing values are problematic for the analysis of microarray data. Imputation methods have been compared in terms of the similarity between imputed and true values in simulation experiments and not of their influence on the final analysis. The focus has been on missing at random, while entries are missing also not at random. RESULTS We investigate the influence of imputation on the detection of differentially expressed genes from cDNA microarray data. We apply ANOVA for microarrays and SAM and look to the differentially expressed genes that are lost because of imputation. We show that this new measure provides useful information that the traditional root mean squared error cannot capture. We also show that the type of missingness matters: imputing 5% missing not at random has the same effect as imputing 10-30% missing at random. We propose a new method for imputation (LinImp), fitting a simple linear model for each channel separately, and compare it with the widely used KNNimpute method. For 10% missing at random, KNNimpute leads to twice as many lost differentially expressed genes as LinImp. AVAILABILITY The R package for LinImp is available at http://folk.uio.no/idasch/imp.

A stochastic model for infectious salmon anemia (ISA) in Atlantic salmon farming

Infectious salmon anemia (ISA) is one of the main infectious diseases in Atlantic salmon farming with major economical implications. Despite the strong regulatory interventions, the ISA epidemic is not under control, worldwide. We study the data covering salmon farming in Norway from 2002 to 2005 and propose a stochastic space-time model for the transmission of the virus. We model seaway transmission between farm sites, transmission through shared management and infrastructure, biomass effects and other potential pathways within the farming industry. We find that biomass has an effect on infectiousness, the local contact network and seaway distance of 5 km represent similar risks, but a large component of risk originates from other sources, among which are possibly infected salmon smolt and boat traffic.

A Bayesian hierarchical model with spatial variable selection: the effect of weather on insurance claims

Climate change will affect the insurance industry. We develop a Bayesian hierarchical statistical approach to explain and predict insurance losses due to weather events at a local geographic scale. The number of weather-related insurance claims is modelled by combining generalized linear models with spatially smoothed variable selection. Using Gibbs sampling and reversible jump Markov chain Monte Carlo methods, this model is fitted on daily weather and insurance data from each of the 319 municipalities which constitute southern and central Norway for the period 1997–2006. Precise out-of-sample predictions validate the model. Our results show interesting regional patterns in the effect of different weather covariates. In addition to being useful for insurance pricing, our model can be used for short-term predictions based on weather forecasts and for long-term predictions based on downscaled climate models.

Comparing methods for handling missing values in food-frequency questionnaires and proposing k nearest neighbours imputation: effects on dietary intake in the Norwegian Women and Cancer study (NOWAC)

OBJECTIVE To investigate item non-response in a postal food-frequency questionnaire (FFQ), and to assess the effect of substituting/imputing missing values on dietary intake levels in the Norwegian Women and Cancer study (NOWAC). We have adapted and probably for the first time applied k nearest neighbours (KNN) imputation to FFQ data. DESIGN Data from a recent reproducibility study were used. The FFQ was mailed twice (test-retest) about 3 months apart to the same subjects. Missing responses in the test FFQ were imputed using the null value (frequencies = null, amount = smallest), the sample mode, the sample median, KNN, and retest values. SETTING A methodological substudy of NOWAC, a national population-based cohort. SUBJECTS A random sample of 2000 women aged 46-75 years was drawn from the cohort in 2002 (response 75%). The imputation methods were compared for 1430 women who completed at least 50% of the test FFQ. RESULTS We imputed 16% missing values in the overall test data matrix. Compared to null value imputation, the largest differences in estimated dietary intake were seen for KNN, and for food items with a high proportion of missing. Imputation with retest values increased total energy intake, indicating that not all missing values are caused by respondents failing to specify no consumption, and that null value imputation may lead to underestimation and misclassification. CONCLUSION Missing values in FFQs present a methodological challenge. We encourage the application and evaluation of newer imputation methods, including KNN, which may reduce imputation errors and give more accurate intake estimates.