Deep learning-based simultaneous bioavailability assessment and speciation analysis of dissolved organic copper

Abstract

Algal copper uptake (i.e., Cu bioavailability) in the euphotic zone plays a vital role in algal photosynthesis and respiration, affecting the primary productivity and the source and sink of atmospheric carbon. Algal Cu uptake is controlled by natural dissolved organic Cu (DOCu) speciation (i.e., complexed with the dissolved organic matter) that conventionally could be tested by model prediction or molecular-level characterizations in the lab, while DOCu uptake are hardly directly assessed. Thus, the new chemistry-biology insight into the mechanisms of the Cu uptake process in algae is urgent. The DOCu speciation transformation (organic DOCu to free Cu(II) ions), enzymatic reduction-induced valence change (reduction of free Cu(II) to Cu(I) ions), and algal Cu uptake at the algae-water interface are imitated. Herein, an intelligent system with DOCu colorimetric sensor is developed for real-time monitoring of newly generated Cu(I) ions. Deep learning with whole sample image-based characterization and powerful feature extraction capabilities facilitates colorimetric measurement. In this context, the Cu bioavailability with 7 kinds of organic ligands (e.g., amino acids, organic acids, carbohydrates) can be predicted by the mimetic intelligent biosensor within 15.0 min, i.e., the DOCu uptake and speciation is successfully predicted and streamlined by the biomimetic approach.