Streams are important sites of transformation of dissolved organic matter (DOM). The molecular characterization of DOM-quality changes requires sophisticated analytical evaluation techniques. The goal of our study was to link molecular DOM transformation with bacterial activity. We measured the degradation of leaf leachate over a gradient of bacterial production obtained by different rates of percolation of sediments in seven experimental flumes on five sampling dates. We developed a new strategy for evaluating molecular formula data sets obtained by ultra-high resolution Fourier transform ion cyclotron resonance mass spectrometry (FTICR MS), in which the time-dependent change of component abundance was fitted by a linear regression model after normalization of mass peak intensities. All components were categorized by calculating the slope (change of percent intensity per day) in each of the seven flumes. These slopes were then related to cumulative bacterial production. The concentration of DOM decreased quickly in all flumes. Bacterial activity was higher in flumes with percolated sediment than in those without percolation, whereas plankton bacterial activity was higher in flumes without percolation or without sediment. There were no differences in molecular-DOM characteristics between flumes, but there were distinct changes over time. Positive slopes, i.e. increasing intensities over time, were found for small molecules (MW < 450 Da) and high O/C ratios, whereas decreasing intensities were observed less often and only for large molecules and low O/C ratios. The positive slopes of produced components showed a positive relationship to bacterial production for small and for oxygen-rich components. The negative slopes of degraded components were negatively related to bacterial production for large and for oxygen-deficient molecules. Overall, the approach provided new insights into the transformation of specific molecular DOM components.