![\"slice](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F177309%2Fca25da59e794a33eec330f878f4c0203%2FScreenshot%20from%202025-03-20%2010-25-27.png?generation=1742480835070239&alt=media\")
Thanks to the host for providing a video about different features of a bacterial cell. https://www.cellstructureatlas.org/6-2-flagellar-motor.html
\nJust need some help identifying the bacterial cell(s) in certain slices. For example, in tomo_6cf2df, the train_labels.csv indicates that there are 2 motors. One in slice 337, and another one in slice 436. However, I am unable to visually locate the bacterial cell (and the motors) on those slices.
\n\n", "messages": [ { "id": 3153585, "postDate": "2025-03-19T01:44:16.103Z", "content": "Thanks to the host for providing a video about different features of a bacterial cell. https://www.cellstructureatlas.org/6-2-flagellar-motor.html
\nJust need some help identifying the bacterial cell(s) in certain slices. For example, in tomo_6cf2df, the train_labels.csv indicates that there are 2 motors. One in slice 337, and another one in slice 436. However, I am unable to visually locate the bacterial cell (and the motors) on those slices.
\n\n", "rawMarkdown": "Thanks to the host for providing a video about different features of a bacterial cell. https://www.cellstructureatlas.org/6-2-flagellar-motor.html\n\nJust need some help identifying the bacterial cell(s) in certain slices. For example, in tomo_6cf2df, the train_labels.csv indicates that there are 2 motors. One in slice 337, and another one in slice 436. However, I am unable to visually locate the bacterial cell (and the motors) on those slices.\n\nSlice 436: \nhttps://storage.googleapis.com/kagglesdsdata/competitions/91249/11294684/train/tomo_6cf2df/slice_0436.jpg?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=databundle-worker-v2%40kaggle-161607.iam.gserviceaccount.com%2F20250319%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20250319T013520Z&X-Goog-Expires=345600&X-Goog-SignedHeaders=host&X-Goog-Signature=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\n\nSlice 337:\nhttps://storage.googleapis.com/kagglesdsdata/competitions/91249/11294684/train/tomo_6cf2df/slice_0337.jpg?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=databundle-worker-v2%40kaggle-161607.iam.gserviceaccount.com%2F20250319%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20250319T013506Z&X-Goog-Expires=345600&X-Goog-SignedHeaders=host&X-Goog-Signature=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", "votes": 6 }, { "id": 3155270, "postDate": "2025-03-20T21:45:59.883Z", "content": "https://www.kaggle.com/code/andrewjdarley/observing-tomo-6cf2df
\nTake a look at the gif generated by this notebook. All I do is concatenate the last frames together and add annotations to represent proximity to the location of the annotation.
\nIf you take a close look, you can see that there is the silhouette of a flagellum pointing to the annotation on slice 337, which is enough to motivate us trusting the annotation there. Wherever there is a flagellum there is a motor. We can see through the annotation on slice 436 that the converse is not true. I watched the gif several times and the only thing I could really see was a bit of a shadow between the inner and outer membranes. I'm a data scientist, not a microbiologist, so I can't justify why there is a motor there. I will just say that in the process of annotating, visual criteria were not the only ones employed. What looks like noise to my untrained eyes has significance whether or not we can make sense of it. The hope is that our models, assuming they don't learn to just look for flagella, can learn to recognize these patterns.
", "rawMarkdown": "https://www.kaggle.com/code/andrewjdarley/observing-tomo-6cf2df\n\nTake a look at the gif generated by this notebook. All I do is concatenate the last frames together and add annotations to represent proximity to the location of the annotation. \n\nIf you take a close look, you can see that there is the silhouette of a flagellum pointing to the annotation on slice 337, which is enough to motivate us trusting the annotation there. Wherever there is a flagellum there is a motor. We can see through the annotation on slice 436 that the converse is not true. I watched the gif several times and the only thing I could really see was a bit of a shadow between the inner and outer membranes. I'm a data scientist, not a microbiologist, so I can't justify why there is a motor there. I will just say that in the process of annotating, visual criteria were not the only ones employed. What looks like noise to my untrained eyes has significance whether or not we can make sense of it. The hope is that our models, assuming they don't learn to just look for flagella, can learn to recognize these patterns. ", "votes": 3 }, { "id": 3154949, "postDate": "2025-03-20T14:44:10.530Z", "content": "Here is my take on the location of the cell in slice 337 from in tomo_6cf2df. In that slice there are two darker lines. Those could correspond to a cell's outer membrane (red arrow) and inner membrane (gold arrow). The image is darker above those two lines, suggesting that area is inside the bacterial cell.
\nThe labeled motor location from the training data is marked with a white arrow. It is a point on the inner cell membrane, so it seems plausible there could be a motor there. However, it isn't obvious to me that there is a motor at that location. I'm not an expert in this, so you might want to take my comment with a grain of salt :)
\n
https://www.kaggle.com/code/andrewjdarley/observing-tomo-6cf2df
\nTake a look at the gif generated by this notebook. All I do is concatenate the last frames together and add annotations to represent proximity to the location of the annotation.
\nIf you take a close look, you can see that there is the silhouette of a flagellum pointing to the annotation on slice 337, which is enough to motivate us trusting the annotation there. Wherever there is a flagellum there is a motor. We can see through the annotation on slice 436 that the converse is not true. I watched the gif several times and the only thing I could really see was a bit of a shadow between the inner and outer membranes. I'm a data scientist, not a microbiologist, so I can't justify why there is a motor there. I will just say that in the process of annotating, visual criteria were not the only ones employed. What looks like noise to my untrained eyes has significance whether or not we can make sense of it. The hope is that our models, assuming they don't learn to just look for flagella, can learn to recognize these patterns.
", "votes": 3, "replies": [] }, { "id": 3154949, "author_name": "riley", "author_url": "", "post_date": "2025-03-20T14:44:10.530000", "content": "Here is my take on the location of the cell in slice 337 from in tomo_6cf2df. In that slice there are two darker lines. Those could correspond to a cell's outer membrane (red arrow) and inner membrane (gold arrow). The image is darker above those two lines, suggesting that area is inside the bacterial cell.
\nThe labeled motor location from the training data is marked with a white arrow. It is a point on the inner cell membrane, so it seems plausible there could be a motor there. However, it isn't obvious to me that there is a motor at that location. I'm not an expert in this, so you might want to take my comment with a grain of salt :)
\n