Under copyright Constraint(s) on Use: This work is protected by the U.S. Copyright Law (Title 17, U.S.C.). Use of this work beyond that allowed by "fair use" or any license applied to this work requires written permission of the copyright holder(s). Responsibility for obtaining permissions and any use and distribution of this work rests exclusively with the user and not the UC San Diego Library. Inquiries can be made to the UC San Diego Library program having custody of the work. Use: This work is available from the UC San Diego Library. This digital copy of the work is intended to support research, teaching, and private study.
Rights Holder and Contact
Salk Institute for Biological Studies
Description
This dataset contains all the movie files, derivative files generated by FlyTracker and JAABA, and all the behavior classifiers described in our article (Leng, X., et al., 2020). In other words, there are raw data which could not be included in the article due to space limitations. With our custom codes (available at https://github.com/asahinak/Leng_Wohl_etal_2020_JAABA_vs_human_annotation), readers can reproduce all the results presented in the article. Contents: There are 3 groups of files. The first contains the classifiers, and the second contains files associated with movies that capture fly behaviors. The third is a single file, ‘featureConfigEyrun.xml’, which is required for JAABA to read outputs of FlyTracker. 1) Classifiers Classifiers for each of lunges, headbutts, and wing extensions are organized into 3 separate directories. The first is “[behavior name]_PartialTraining” directory, the second is “[behavior name]_Downsampled” directory, and the third is “[behavior name]_SelectFeatures” directory. A PartialTraining directory contains 4 .JAB classifiers that were trained with only a subset of the complete training frames, along with a classifier trained with all available frames (total 5 classifiers). The number at the end of the file name corresponds to the number of classifiers described in Supplementary Table S6. See Fig. 4 of the article for the performance evaluations of these classifiers, and Supplementary Tables S1-S4 for details of training frames. A Downsampled directory contains 50 .JAB classifiers that were trained with downsampled training frames. Ten classifiers trained with randomly chosen training frames (described in details in Supplementary Table S6 the article) were generated for each of 5 downsampling rates. See Fig. 4 and Supplementary Fig. S4 of the article for the summary of downsampled classifiers’ performance. A SelectFeatures directory contains 8 .JAB classifiers that were trained with selected features or rules. See Fig. 5 and Supplementary Fig. S5 of the article for the performance evaluations of these classifiers, and Supplementary Table S5 of the article for the details of features or rules that were selectively eliminated. 2) Movies and associated files One .AVI movie and associated files are grouped into one .ZIP file. See Table 1 and Supplementary Table S1 of the article for detailed metadata, such as genotypes of the flies, experimental conditions, and fly pairs used for classifier training and performance evaluation. The extracted file has a nested directory structure along with various files in each subdirectory. It is important to maintain the directory structure for FlyTracker and JAABA to operate properly. For details of these programs, please see the following link and documentation: FlyTracker: http://www.vision.caltech.edu/Tools/FlyTracker/ (Reference: Eyjolfsdottir, E. et al. (2014) Detecting social actions of fruit flies. In: Fleet D., Pajdla T., Schiele B., Tuytelaars T. (eds) Computer Vision – ECCV 2014. ECCV 2014. Lecture Notes in Computer Science, vol 8690. Springer, Cham. https://doi.org/10.1007/978-3-319-10605-2_50) JAABA: http://jaaba.sourceforge.net/ (Reference: Kabra M. et al., JAABA: interactive machine learning for automatic annotation of animal behavior. Nat Methods 10, 64–67 (2013). https://doi.org/10.1038/nmeth.2281) The base directory contains the following 3 components: - An .AVI movie ([movie_name].avi): an original movie of behaving fly pairs. - calibration.mat file: created during calibration step of FlyTracker. - A [movie_name] subdirectory The [movie_name] subdirectory contains the following 4 components: - A [movie_name]-bg.mat file: a background file created by FlyTracker. - A [movie_name]-feat.mat file: a feature file created by FlyTracker - A [movie_name]-track.mat file: a tracking file created by FlyTracker - A [movie_name]_JAABA subdirectory The [movie_name]_JAABA subdirectory contains the following 2 components: - A trx.mat file: JAABA’s tracking file - A ‘perframe’ subdirectory The perframe subdirectory contains 39 JAABA features (.mat) as described in Eyjolfsdottir et al. (see above) and Fig. 2 of the article. Note that we removed a “movie.avi” file under the [movie_name]_JAABA subdirectory to reduce the file size. Copy the original movie into this subdirectory, and rename the file to “movie.avi” if you wish to visualize the movie in the JAABA training mode. 3) featureConfigEyrun.xml This file is kindly provided by Eyrun Eyjolfsdottir in Prof. Pietro Perona’s lab at Caltech. It is necessary for JAABA to create a [movie_name]_JAABA subdirectory with correct perframe feature files from output files of FlyTracker. When JAABA is installed, copy the featureConfigEyrun.xml file in the ‘JAABA-masterperframeparams’ directory. FlyTracker output files should be readable by JAABA. NIGMS R35 GM119844 (to K.A.), the Japan Society for the Promotion of Science (to K.I.), the Naito Foundation (to K.I.), the Mary K. Chapman Foundation (to M.W.), the Rose Hills Foundation (to M.W.) Research Data Curation Program, UC San Diego, La Jolla, 92093-0175 (https://lib.ucsd.edu/rdcp) Leng, Xubo; Wohl, Margot; Ishii, Kenichi; Nayak, Pavan; Asahina, Kenta (2020). Data from: Quantifying influence of human choice on the automated detection of Drosophila behavior by a supervised machine learning algorithm. UC San Diego Library Digital Collections. https://doi.org/10.6075/J0QF8RDZ Is Supplement To: Leng, X., Wohl, M., Ishii, K., Nayak, P., Asahina, K. Quantifying influence of human choice on the automated detection of Drosophila behavior by a supervised machine learning algorithm. PLoS ONE. https://doi.org/10.1371/journal.pone.0241696 Is Referenced By: Ishii, K., Wohl, M., DeSouza, A., Asahina, K. Sex-determining genes distinctly regulate courtship capability and target preference via sexually dimorphic neurons. eLife 9, e52701 (2020). https://doi.org/10.7554/eLife.52701 Wohl, M., Ishii, K., Asahina, K. Layered roles of fruitless isoforms in specification and function of male aggression-promoting neurons in Drosophila. eLife 9, e52702 (2020). https://doi.org/10.7554/eLife.52702 References: Eyjolfsdottir, E. et al. (2014) Detecting social actions of fruit flies. In: Fleet D., Pajdla T., Schiele B., Tuytelaars T. (eds) Computer Vision – ECCV 2014. ECCV 2014. Lecture Notes in Computer Science, vol 8690. Springer, Cham. https://doi.org/10.1007/978-3-319-10605-2_50 Kabra M. et al., JAABA: interactive machine learning for automatic annotation of animal behavior. Nat Methods 10, 64–67 (2013). https://doi.org/10.1038/nmeth.2281
Type
Dataset
Language
English
Subject
Machine vision Behavior classification Social behavior Sexual behavior Machine learning Drosophila melanogaster
If you're wondering about permissions and what you can do with this item, a good starting point is the "rights information" on this page. See our terms of use for more tips.
Share your story
Has Calisphere helped you advance your research, complete a project, or find something meaningful? We'd love to hear about it; please send us a message.
