When the wrong protein wins: hidden contaminants in cryo-EM

What is it about?

This study highlights an unexpected but important problem in structural biology: proteins from the host organism can contaminate purified samples and mislead cryo-electron microscopy (cryo-EM) analysis. While attempting to determine the structure of a motor protein complex (KIF17–IFT70), we instead obtained a high-resolution structure of an unrelated bacterial protein called ArnA. Although the sample appeared clean by standard biochemical methods such as SDS–PAGE and mass spectrometry, the contaminant remained hidden and ultimately dominated the cryo-EM dataset. We explain why such contaminants are difficult to detect using conventional quality-control approaches and show how cryo-EM can reveal them. This work emphasizes the need for improved validation strategies and early screening to ensure that researchers are studying their intended targets.

Why is it important?

This work addresses a growing and underappreciated challenge in cryo-EM: the risk that biologically irrelevant contaminants can dominate datasets and lead to unintended structures. Unlike traditional structural methods, cryo-EM is particularly sensitive to particle homogeneity and alignment, allowing stable contaminants such as ArnA to outcompete the intended target—even when present at low levels. Our findings show that standard biochemical quality-control techniques may fail to detect such contaminants, especially for low-yield or flexible protein complexes. By systematically analyzing this discrepancy, we provide practical recommendations for improving purification and screening workflows. This is timely as cryo-EM becomes a mainstream method, and avoiding misinterpretation is critical for both basic research and drug discovery.