In Heimowitz et al. 2018 we introduced the APPLE-Picker, a template-free and training-free, fast, and accurate computational framework for automatic particle picking. We released the APPLE picker as a stand-alone package in both Python and Matlab on Github. The APPLE-Picker takes 7 seconds per micrograph on a single GPU (Nvidia Tesla P100). On a CPU (2.6 GHz Intel core i7 cpu with 4 cores and 16 GB of memory), the Matlab version takes approximately 2 minutes for each micrograph. The Python version takes 67 seconds per micrograph when processing sequentially. Micrographs can be processed in parallel on multiple cores. The APPLE-Picker was evaluated on many publicly available datasets and high-resolution reconstructions directly from the automatically picked particles without any manual intervention are routinely obtained, for example, a 4.5 Å β-galactosidase and 3.4 Å T20S proteasome.
A. Heimowitz, J. Andén, A. Singer, APPLE Picker: Automatic Particle Picking, a Low-Effort Cryo-EM Framework, Journal of Structural Biology, 204(2), pp. 215-227, 2018. PMID: 30134153 [PubMed] PMCID: PMC6183064.