Validity of machine learning in biology and medicine increased through collaborations across fields of expertise

Abstract

Machine Learning (ML) has become an essential asset for the life sciences and medicine. We selected 300 articles describing ML applications from 17 journals sampling 26 different fields between 2011 and 2018. Independent evaluation by two readers highlighted three results.First, only half of the articles shared software, 64% shared data, and 81% applied any kind of evaluation. Although these aspects are crucial to ensure validity and reliability of ML applications, they were met more by publications in lower-ranked journals. Second, the authors' scientific background highly influenced how technical aspects were addressed: reproducibility and computational evaluation methods were more prominent with computational co-authors; experimental proofs more with experimentalists. Third, 73% of the ML applications resulted from interdisciplinary collaborations comprising authors from at least two of the three disciplines: computational sciences, experimental biology, medicine._deleted_ The data suggested collaborations between computational and experimental scientists to generate more computationally sound and impactful work integrating knowledge.Furthermore, such collaborations provide opportunities to both sides: computational scientists are given access to novel and challenging real-world biological data increasing the scientific impact of their research, and experimentalists benefit from more in-depth computational analyses improving the technical correctness of work.

Integration of mechanistic immunological knowledge into a machine learning pipeline improves predictions

“…A structured collaboration between clinicians, biologists and computer scientists can lead to machine learning algorithms in the life sciences that achieve stronger results 67 . In this Article, we proposed a collaborative framework that enables integration of prior knowledge of cell signalling pathways in a machine learning algorithm to improve the predictions and robustness of the resulting models in clinical datasets.…”

New Therapeutic Targets in Autoimmune Cholangiopathies

“…It is key the improvement of risk stratification strategies for this purpose (98), which will require deep, longitudinal phenotyping of individuals by means of multi-omics analysis including the exposome, along with the microbiome, genome, metabolome, among the others (99)(100)(101)(102)(103)(104). Then, we need to implement algorithms to proficiently integrate these big data to cluster patients across different phenotypes and trajectories of the disease (105,106); for this, the collaboration with data science professionals and experts in artificial intelligence will be fundamental (107,108). The last part of this process will be to put in practice clinical trials with different, multimodal treatment strategies (109).…”

Integration of Mechanistic Immunological Knowledge into a Machine Learning Pipeline Increases Predictive Power

“…A structured collaboration between clinicians, biologists, and computer scientists can lead to machine learning algorithms in life sciences achieving stronger results 53 . In this article, we proposed a collaborative framework that enables integration of prior knowledge of cell signaling pathways in a machine learning algorithm to increase the predictive power and robustness of the resulting models in clinical datasets.…”