Traces from single-molecule fluorescence microscopy (SMFM) experiments exhibit photophysical artifacts that typically necessitate human expert screening, which is time-consuming and introduces potential for user-dependent expectation bias. Here, we have used deep learning to develop a rapid, automatic SMFM trace selector, termed AutoSiM, that improves the sensitivity and specificity of an assay for a DNA point mutation based on single-molecule recognition through equilibrium Poisson sampling (SiMREPS). The improved performance of AutoSiM is based on accepting both more true positives and fewer false positives than the conventional approach of hidden Markov modeling (HMM) followed by thresholding. As a second application, the selector was used for automated screening of single-molecule Förster resonance energy transfer (smFRET) data to identify high-quality traces for further analysis, and achieves ~90% concordance with manual selection while requiring less processing time. AutoSiM can be adapted readily to novel datasets, requiring only modest Transfer Learning.
Li, J., Zhang, L., Johnson-Buck, A., & Walter, N. G. (2020). Automatic classification and segmentation of single-molecule fluorescence time traces with deep learning. Nature Communications, 11(1), 5833. https://doi.org/10.1038/s41467-020-19673-1 and Hayward, S., Lund, P., Kang, Q., Johnson-Buck, A., Tewari, M., Walter, N. (2018). Single-molecule microscopy image data and analysis files for "Ultra-specific and Amplification-free Quantification of Mutant DNA by Single-molecule Kinetic Fingerprinting" [Data set]. University of Michigan - Deep Blue. https://doi.org/10.7302/Z2CZ35DF
The data were used to study the high-frequency geomagnetic disturbances within the magnetic field data. Included in this repository are the python scripts that perform an identification and classification of high-frequency signals within the magnetometer data that is downloaded from the databases listed in the Methodology section. All analysis and plots were created using subsequent Python libraries. The machine learning study implemented libraries from the sci-kit learn software. All of the specific methodology can be accessed in the readme.txt script.
The work on accelerating authenticated boot for embedded system resulted in designing an algorithm in python to perform the random address generation and cryptographic MAC calculation.
The Sampled Boot schemes implemented in this package allow a significant reduction of the time
needed to authenticate firmware images during startup, while still retaining a high degree of trust.
This is particularly useful for automotive applications in which startup time constraints make secure boot a time prohibitive process. and Citation for this dataset: Nasser, A., Gumise, W. (2019). Authenticated Boot Acceleration Algorithm [Code and data]. University of Michigan Deep Blue Data Repository. https://doi.org/10.7302/yeh1-1x17
Nasser, A., Gumise, W., and Ma, D., "Accelerated Secure Boot for Real-Time Embedded Safety Systems," SAE Int. J. Transp. Cyber. & Privacy 2(1) : 35-48, 2019, https://doi.org/10.4271/11-02-01-0003
recordings made in Barato village. Referred to as "text 2021-02" and "text 2021-03." Text 2021-03 is transcribed and annotated at the end of the reference grammar (see link to Deep Blue Documents). Text 2021-02 covers a subset of the same content and has not been transcribed as of late 2022. See also "notes" file inside the work.
A subset of the Kelenga recordings are being transcribed and will serve as data for the Kelenga reference grammar which, when finished, will be included in the collection "Bozo languages of Mali (documents)" in Deep Blue Documents (see link).
these and other recordings are data for a reference grammar of Kelenga that, when completed, will be archived in the collection "Bozo languages of Mali (documents)" in Deep Blue Documents. For contents see the "notes" file inside the work. A few of the Kelenga texts are being transcribed, others will be left for others to transcribe or listen to as they wish.
For content see the "notes" file inside the work. Most of the recordings are translated and annotated at the end of the reference grammar (see link to Deep Blue Documents).