Binary classifiers are routinely evaluated with performance measures such as sensitivity and specificity, and performance is frequently illustrated with Receiver Operating Characteristics (ROC) plots. Alternative measures such as positive predictive value (PPV) and the associated Precision/Recall (PRC) plots are used less frequently. Many bioinformatics studies develop and evaluate classifiers that are to be applied to strongly imbalanced datasets in which the number of negatives outweighs the number of positives significantly. While ROC plots are visually appealing and provide an overview of a classifier's performance across a wide range of specificities, one can ask whether ROC plots could be misleading when applied in imbalanced classification scenarios. We show here that the visual interpretability of ROC plots in the context of imbalanced datasets can be deceptive with respect to conclusions about the reliability of classification performance, owing to an intuitive but wrong interpretation of specificity. PRC plots, on the other hand, can provide the viewer with an accurate prediction of future classification performance due to the fact that they evaluate the fraction of true positives among positive predictions. Our findings have potential implications for the interpretation of a large number of studies that use ROC plots on imbalanced datasets.

SummaryThe precision–recall plot is more informative than the ROC plot when evaluating classifiers on imbalanced datasets, but fast and accurate curve calculation tools for precision–recall plots are currently not available. We have developed Precrec, an R library that aims to overcome this limitation of the plot. Our tool provides fast and accurate precision–recall calculations together with multiple functionalities that work efficiently under different conditions.Availability and ImplementationPrecrec is licensed under GPL-3 and freely available from CRAN (https://cran.r-project.org/package=precrec). It is implemented in R with C ++.Supplementary information Supplementary data are available at Bioinformatics online.

MicroRNAs (miRNAs) are a class of small noncoding RNAs that can regulate many genes by base pairing to sites in mRNAs. The functionality of miRNAs overlaps that of short interfering RNAs (siRNAs), and many features of miRNA targeting have been revealed experimentally by studying miRNA-mimicking siRNAs. This review outlines the features associated with animal miRNA targeting and describes currently available prediction tools.

MicroRNAs (miRNAs) regulate genes post transcription by pairing with messenger RNA (mRNA). Variants such as single nucleotide polymorphisms (SNPs) in miRNA regulatory regions might result in altered protein levels and disease. Genome-wide association studies (GWAS) aim at identifying genomic regions that contain variants associated with disease, but lack tools for finding causative variants. We present a computational tool that can help identifying SNPs associated with diseases, by focusing on SNPs affecting miRNA-regulation of genes. The tool predicts the effects of SNPs in miRNA target sites and uses linkage disequilibrium to map these miRNA-related variants to SNPs of interest in GWAS. We compared our predicted SNP effects in miRNA target sites with measured SNP effects from allelic imbalance sequencing. Our predictions fit measured effects better than effects based on differences in free energy or differences of TargetScan context scores. We also used our tool to analyse data from published breast cancer and Parkinson's disease GWAS and significant trait-associated SNPs from the NHGRI GWAS Catalog. A database of predicted SNP effects is available at http://www.bigr.medisin.ntnu.no/mirsnpscore/. The database is based on haplotype data from the CEU HapMap population and miRNAs from miRBase 16.0.

Feeding plant‐based diet through smoltification of Atlantic salmon requires verification of the optimal level of 1C nutrients. Here, we fed Atlantic salmon plant‐based diets containing three different surplus amounts of the 1C nutrients; methionine, cobalamin (vitamin B12), pyridoxine (vitamin B6) and folic acid during 6 weeks in fresh water, through smoltification, followed by 3 months on‐growing period in salt water. The three diets were fed to fish dispersed in triplicate tanks throughout the experiment. Mean start body weight was 32 g. Dietary methionine levels in the diets were 6.7, 9.2 and 11.7 g/kg. Dietary B6 was 6.75, 8.45 and 11 mg/kg. Cobalamin was 0.16, 0.18 and 0.20 mg/kg. While dietary folic acid was 2.9, 4.8 and 6.3 mg/kg, diets are referred to as low, medium and high 1C diet. All other amino acids were similar between diets. The results showed no differences in growth or feed utilization in the fresh water period, but following the on‐growing salt water period, differences between diets occurred. The fish fed the medium 1C diet showed better growth, as compared to fish fed the low or high 1C diet (p = .009). The medium 1C fed fish showed a relative lower liver weight compared with fish fed low or high 1C diet (p = .025). Condition factor was better in fish fed the medium and high 1C diet as compared to those fed the low 1C diet (p = .0006). As expected, free methionine in liver, plasma and muscle increased by dietary methionine inclusion. Surplus vitamins only had minor effect on tissue concentrations. Based on these findings, we conclude that the micronutrient and methionine level presented in the medium 1C diet improved the growth, liver size and condition factor; however, more research is needed to evaluate the optimal requirement level for each of the 1C nutrients.