Best practices for using biological data in psychological research

Advances in statistical techniques for analyzing biological data have driven the need for larger sample sizes, which are ultimately made possible through team science efforts. Findings from these efforts have led to a significant shift in approaches to biological data collection over the past decade, from traditional approaches relying on single measures or assays collected by individual investigators in small lab-based studies to high-dimensional methodologies embedded within large cohorts or multi-site studies. These recent large scale consortium efforts necessitate careful consideration across the full data pipeline, from pre-data generation and sample collection to processing and statistical analysis, while addressing challenges such as cost, standardization of laboratory procedures, and harmonization across sites. Selecting an appropriate methodological approach to measure a biomarker involves complex decisions and tradeoffs to determine the most suitable biological modality (e.g., blood vs. saliva) and assay type (e.g., plasma vs. serum) for the research question and target population. These decisions must account for contextual factors such as cost, study population characteristics, storage space, and analytic capabilities, and carry important implications for data quality, feasibility, and interpretability. Standardization of data collection, storage, quality control, and analytical practices (e.g., accounting for technical confounders), is critical, especially when analyzing biological data across multiple cohorts and time points. This review synthesizes recommended best practices for standardizing collection of peripheral biomarkers (blood, saliva, and urine) in psychiatric research, which we hope will enhance data quality, facilitate cross-study comparisons, and advance individual and team biomarker research now and in the future.