BACKGROUND: Numerous transcriptomic-based models have been developed to predict or understand the fundamental mechanisms driving biological phenotypes. However, few models have successfully transitioned into clinical practice due to challenges associated with generalizability and interpretability. To address these issues, researchers have turned to dimensionality reduction methods and have begun implementing transfer learning approaches. METHODS: In this study, we aimed to determine the optimal combination of dimensionality reduction and regularization methods for predictive modeling. We applied seven dimensionality reduction methods to various datasets, including two supervised methods (linear optimal low-rank projection and low-rank canonical correlation analysis), two unsupervised methods [principal component analysis and consensus independent component analysis (c-ICA)], and three methods [autoencoder (AE), adversarial variational autoencoder, and c-ICA] within a transfer learning framework, trained on > 140,000 transcriptomic profiles. To assess the performance of the different combinations, we used a cross-validation setup encapsulated within a permutation testing framework, analyzing 30 different transcriptomic datasets with binary phenotypes. Furthermore, we included datasets with small sample sizes and phenotypes of varying degrees of predictability, and we employed independent datasets for validation. RESULTS: Our findings revealed that regularized models without dimensionality reduction achieved the highest predictive performance, challenging the necessity of dimensionality reduction when the primary goal is to achieve optimal predictive performance. However, models using AE and c-ICA with transfer learning for dimensionality reduction showed comparable performance, with enhanced interpretability and robustness of predictors, compared to models using non-dimensionality-reduced data. CONCLUSION: These findings offer valuable insights into the optimal combination of strategies for enhancing the predictive performance, interpretability, and generalizability of transcriptomic-based models.

MOTIVATION: Sample mix-ups can arise during sample collection, handling, genotyping or data management. It is unclear how often sample mix-ups occur in genome-wide studies, as there currently are no post hoc methods that can identify these mix-ups in unrelated samples. We have therefore developed an algorithm (MixupMapper) that can both detect and correct sample mix-ups in genome-wide studies that study gene expression levels. RESULTS: We applied MixupMapper to five publicly available human genetical genomics datasets. On average, 3% of all analyzed samples had been assigned incorrect expression phenotypes: in one of the datasets 23% of the samples had incorrect expression phenotypes. The consequences of sample mix-ups are substantial: when we corrected these sample mix-ups, we identified on average 15% more significant cis-expression quantitative trait loci (cis-eQTLs). In one dataset, we identified three times as many significant cis-eQTLs after correction. Furthermore, we show through simulations that sample mix-ups can lead to an underestimation of the explained heritability of complex traits in genome-wide association datasets. AVAILABILITY AND IMPLEMENTATION: MixupMapper is freely available at http://www.genenetwork.nl/mixupmapper/