Fig. 1: Overview of data extraction, evaluation, and analysis using Deep-LASI. | Nature Communications

Fig. 1: Overview of data extraction, evaluation, and analysis using Deep-LASI.

From: Deep-LASI: deep-learning assisted, single-molecule imaging analysis of multi-color DNA origami structures

Fig. 1

a Single-molecule data of up to three separate channels after direct and alternating laser excitation are identified, extracted, and presorted for further analyses. Each frame within the time traces is classified into categories using a hybrid CNN-LSTM. b A second hybrid CNN-LSTM evaluates the kinetics and state information in the presorted data. The photobleaching information can be used for determining the correction factors to obtain accurate FRET values between two and three fluorophores. c Next, the interconversion rates between underlying states and absolute, distance-related FRET values are extracted from multi-color datasets.

Back to article page