Author Correction: Analysis of the Human Protein Atlas Image Classification competition.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Publisher Correction: Analysis of the Human Protein Atlas Image Classification competition.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Publisher Correction: Analysis of the Human Protein Atlas Image Classification competition.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Analysis of the Human Protein Atlas Image Classification competition.

Pinpointing subcellular protein localizations from microscopy images is easy to the trained eye, but challenging to automate. Based on the Human Protein Atlas image collection, we held a competition to identify deep learning solutions to solve this task. Challenges included training on highly imbalanced classes and predicting multiple labels per image. Over 3 months, 2,172 teams participated. Despite convergence on popular networks and training techniques, there was considerable variety among the solutions. Participants applied strategies for modifying neural networks and loss functions, augmenting data and using pretrained networks. The winning models far outperformed our previous effort at multi-label classification of protein localization patterns by ~20%. These models can be used as classifiers to annotate new images, feature extractors to measure pattern similarity or pretrained networks for a wide range of biological applications.

Multicenter study of acetaminophen hepatotoxicity reveals the importance of biological endpoints in genomic analyses.

Gene expression profiling is a widely used technique with data from the majority of published microarray studies being publicly available. These data are being used for meta-analyses and in silico discovery; however, the comparability of toxicogenomic data generated in multiple laboratories has not been critically evaluated. Using the power of prospective multilaboratory investigations, seven centers individually conducted a common toxicogenomics experiment designed to advance understanding of molecular pathways perturbed in liver by an acute toxic dose of N-acetyl-p-aminophenol (APAP) and to uncover reproducible genomic signatures of APAP-induced toxicity. The nonhepatotoxic APAP isomer N-acetyl-m-aminophenol was used to identify gene expression changes unique to APAP. Our data show that c-Myc is induced by APAP and that c-Myc-centered interactomes are the most significant networks of proteins associated with liver injury. Furthermore, sources of error and data variability among Centers and methods to accommodate this variability were identified by coupling gene expression with extensive toxicological evaluation of the toxic responses. We show that phenotypic anchoring of gene expression data is required for biologically meaningful analysis of toxicogenomic experiments.