[3dem] Topaz v0.1.0 release with GUI

[Alex Noble]

Dear all,

We are pleased to release Topaz v0.1.0.  Topaz is a novel cryoEM particle picker using deep learning convolutional neural networks within a positive-unlabeled framework:

https://github.com/tbepler/topaz

In our experience, Topaz is able to pick all orientations of any size and shape particle, to avoid aggregation, to avoid junk and grid substrate, to identify more particles than other methods, and to properly order particles by likelihood. These advances: 1) enable conventionally difficult single particle projects to move forward; 2) significantly decrease ad-hoc user post-processing (e.g. particle filtering/2D classification); 3) make classification more robust and representative given the significantly larger number of real particles picked; and 4) increase collection and processing efficiency. These achievements are accomplished by requiring training on only a minimal set of sparse positive particles.

Topaz may be installed on all operating systems using a variety of streamlined methods:

Anaconda (Linux, Mac, Windows):  conda install topaz -c tbepler -c soumith
Docker (Linux, Mac, Windows):  docker pull semc/topaz
Singularity (Linux, Mac, Windows):  singularity pull shub://dallakyan/topaz_singularity
SBGrid (Linux):  https://sbgrid.org/software/titles/topaz
>From source (Linux):  Follow Github instructions<https://github.com/tbepler/topaz/tree/gui#installation>

Topaz is open source under GPLv3, is modular, and thus may be integrated into any cryoEM software. In addition, we provide a standalone GUI with Topaz<https://github.com/tbepler/topaz/tree/master/topaz/gui> based entirely on HTML/JavaScript/CSS and thus functional on all modern web browsers. The GUI may be tested here:

http://emgweb.nysbc.org/topaz.html

A presentation of Topaz’s method and some of its capabilities may be seen here:

https://www.youtube.com/watch?v=I6PUxeVBQt8

Please report any issues, suggestions, or requests as a GitHub issue<https://github.com/tbepler/topaz/issues>.
We hope this software is useful to the cryoEM community.


Best regards,
Tristan Bepler, Computer Science & Artificial Intelligence Laboratory @ MIT, tbepler at mit.edu<mailto:tbepler at mit.edu>
Alex Noble, Simons Electron Microscopy Center @ NYSBC, anoble at nysbc.org<mailto:anoble at nysbc.org>, twitter<https://twitter.com/alexjamesnoble>
Bonnie Berger, Computer Science & Artificial Intelligence Laboratory @ MIT, bab at mit.edu<mailto:bab at mit.edu>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ncmir.ucsd.edu/pipermail/3dem/attachments/20190325/0ee04128/attachment-0001.html>