Magdalena Kircher

Outliers in the training or test set used to fit and evaluate a classifier on transcriptomics data can considerably change the estimated performance of the model. Hence, an either too weak or a too optimistic accuracy is then reported and the estimated model performance cannot be reproduced on independent data. It is then also doubtful whether a classifier qualifies for clinical usage. We estimate classifier performances in simulated gene expression data with artificial outliers and in two real-world datasets. As a new approach, we use two outlier detection methods within a bootstrap procedure to estimate the outlier probability for each sample and evaluate classifiers before and after outlier removal by means of cross-validation. We found that the removal of outliers changed the classification performance notably. For the most part, removing outliers improved the classification results. Taking into account the fact that there are various, sometimes unclear reasons for a sample to be an outlier, we strongly advocate to always report the performance of a transcriptomics classifier with and without outliers in training and test data. This provides a more diverse picture of a classifier’s performance and prevents reporting models that later turn out to be not applicable for clinical diagnoses.

To better understand the molecular basis of respiratory diseases of viral origin, high-throughput gene-expression data are frequently taken by means of DNA microarray or RNA-seq technology. Such data can also be useful to classify infected individuals by molecular signatures in the form of machine-learning models with genes as predictor variables. Early diagnosis of patients by molecular signatures could also contribute to better treatments. An approach that has rarely been considered for machine-learning models in the context of transcriptomics is data augmentation. For other data types it has been shown that augmentation can improve classification accuracy and prevent overfitting. Here, we compare three strategies for data augmentation of DNA microarray and RNA-seq data from two selected studies on respiratory diseases of viral origin. The first study involves samples of patients with either viral or bacterial origin of the respiratory disease, the second study involves patients with either SARS-CoV-2 or another respiratory virus as disease origin. Specifically, we reanalyze these public datasets to study whether patient classification by transcriptomic signatures can be improved when adding artificial data for training of the machine-learning models. Our comparison reveals that augmentation of transcriptomic data can improve the classification accuracy and that fewer genes are necessary as explanatory variables in the final models. We also report genes from our signatures that overlap with signatures presented in the original publications of our example data. Due to strict selection criteria, the molecular role of these genes in the context of respiratory infectious diseases is underlined.

Estimating the taxonomic composition of viral sequences in a biological samples processed by next-generation sequencing is an important step in comparative metagenomics. Mapping sequencing reads against a database of known viral reference genomes, however, fails to classify reads from novel viruses whose reference sequences are not yet available in public databases. Instead of a mapping approach, and in order to classify sequencing reads at least to a taxonomic level, the performance of artificial neural networks and other machine learning models was studied. Taxonomic and genomic data from the NCBI database were used to sample labelled sequencing reads as training data. The fitted neural network was applied to classify unlabelled reads of simulated and real-world test sets. Additional auxiliary test sets of labelled reads were used to estimate the conditional class probabilities, and to correct the prior estimation of the taxonomic distribution in the actual test set. Among the taxonomic levels, the biological order of viruses provided the most comprehensive data base to generate training data. The prediction accuracy of the artificial neural network to classify test reads to their viral order was considerably higher than that of a random classification. Posterior estimation of taxa frequencies could correct the primary classification results.