When considering my observations, keep in mind that I am not a cell biologist - I only play one on TV.
\n\nConsider this when creating image masks:
\n\n1) nuclei will be stained in most images, and the rest of cell will not be stained. Most of the time you will see is bunch of nuclei but we should draw an imaginary line for cell membrane that is around the nucleus. In some images that cell membrane line will be visible.
\n\n2) Nuclei are not uniformly stained, and the granularity of staining depends both on the type of cell and the exact staining method used. In particular, sometimes there will be 1-2 dots inside a nucleus that are nucleoli (singular is nucleolus). Nucleoli can be unstained while the rest of nucleus is stained, though sometimes the opposite is true.
\n\n3) Nuclei of same cells have similar shapes and sizes, though they may seem misshapen at times because of the way cells are laying on the surface. At the same time there are imaging artifacts - sometimes literally specks of dust - that will give strong signal, but are not part of cells. A simple way to exclude those would be to calculate an average area of nuclei and exclude the signal that is smaller than it by some ratio. Also, \"dots\" that have much stronger signal than the rest (5-10x) are almost never cells. Finally, cells also tend to clump together and there should be a way to split three nuclei of tightly packed cells even when there is no obvious space between them.
\n\n4) Sometimes it is the rest of cell - the cytoplasm - that is stained rather than the nucleus, or in addition to it. For example, in image cb4df20a83b2f38b394c67f1d9d4aef29f9794d5345da3576318374ec3a11490 the nucleus is a relatively unstained part in the middle rather than strongly stained dots around it.
\n\n5) Images have different magnification: some of them are at 20-50x, others ate 200-400x. Needless to say, it is much easier to define nuclei precisely at higher magnification.
", "messages": [ { "id": "270043", "postDate": "01/17/2018 17:23:44", "content": "When considering my observations, keep in mind that I am not a cell biologist - I only play one on TV.
\n\nConsider this when creating image masks:
\n\n1) nuclei will be stained in most images, and the rest of cell will not be stained. Most of the time you will see is bunch of nuclei but we should draw an imaginary line for cell membrane that is around the nucleus. In some images that cell membrane line will be visible.
\n\n2) Nuclei are not uniformly stained, and the granularity of staining depends both on the type of cell and the exact staining method used. In particular, sometimes there will be 1-2 dots inside a nucleus that are nucleoli (singular is nucleolus). Nucleoli can be unstained while the rest of nucleus is stained, though sometimes the opposite is true.
\n\n3) Nuclei of same cells have similar shapes and sizes, though they may seem misshapen at times because of the way cells are laying on the surface. At the same time there are imaging artifacts - sometimes literally specks of dust - that will give strong signal, but are not part of cells. A simple way to exclude those would be to calculate an average area of nuclei and exclude the signal that is smaller than it by some ratio. Also, \"dots\" that have much stronger signal than the rest (5-10x) are almost never cells. Finally, cells also tend to clump together and there should be a way to split three nuclei of tightly packed cells even when there is no obvious space between them.
\n\n4) Sometimes it is the rest of cell - the cytoplasm - that is stained rather than the nucleus, or in addition to it. For example, in image cb4df20a83b2f38b394c67f1d9d4aef29f9794d5345da3576318374ec3a11490 the nucleus is a relatively unstained part in the middle rather than strongly stained dots around it.
\n\n5) Images have different magnification: some of them are at 20-50x, others ate 200-400x. Needless to say, it is much easier to define nuclei precisely at higher magnification.
", "rawMarkdown": "When considering my observations, keep in mind that I am not a cell biologist - I only play one on TV.\n\nConsider this when creating image masks:\n\n1) nuclei will be stained in most images, and the rest of cell will not be stained. Most of the time you will see is bunch of nuclei but we should draw an imaginary line for cell membrane that is around the nucleus. In some images that cell membrane line will be visible.\n\n2) Nuclei are not uniformly stained, and the granularity of staining depends both on the type of cell and the exact staining method used. In particular, sometimes there will be 1-2 dots inside a nucleus that are nucleoli (singular is nucleolus). Nucleoli can be unstained while the rest of nucleus is stained, though sometimes the opposite is true.\n\n3) Nuclei of same cells have similar shapes and sizes, though they may seem misshapen at times because of the way cells are laying on the surface. At the same time there are imaging artifacts - sometimes literally specks of dust - that will give strong signal, but are not part of cells. A simple way to exclude those would be to calculate an average area of nuclei and exclude the signal that is smaller than it by some ratio. Also, \"dots\" that have much stronger signal than the rest (5-10x) are almost never cells. Finally, cells also tend to clump together and there should be a way to split three nuclei of tightly packed cells even when there is no obvious space between them.\n\n4) Sometimes it is the rest of cell - the cytoplasm - that is stained rather than the nucleus, or in addition to it. For example, in image cb4df20a83b2f38b394c67f1d9d4aef29f9794d5345da3576318374ec3a11490 the nucleus is a relatively unstained part in the middle rather than strongly stained dots around it.\n\n5) Images have different magnification: some of them are at 20-50x, others ate 200-400x. Needless to say, it is much easier to define nuclei precisely at higher magnification.", "votes": null }, { "id": "270060", "postDate": "01/17/2018 17:50:37", "content": "Several things I think are important as starting points: 1) define/guess image magnification; 2) define what parts are cells and what are imaging artefacts; 3) figure out whether the nucleus is stained or not; 4) figure out whether the nucleus is the only thing that's stained, as sometimes whole cells are part of the signal.
", "rawMarkdown": "Several things I think are important as starting points: 1) define/guess image magnification; 2) define what parts are cells and what are imaging artefacts; 3) figure out whether the nucleus is stained or not; 4) figure out whether the nucleus is the only thing that's stained, as sometimes whole cells are part of the signal.", "votes": null }, { "id": "272318", "postDate": "01/22/2018 20:27:09", "content": "Thank you so much for sharing this information. I was clueless as how to find/interpret the nuclei.
", "rawMarkdown": "Thank you so much for sharing this information. I was clueless as how to find/interpret the nuclei.", "votes": null } ], "comments": [ { "id": 270060, "author_name": "tilii7", "author_url": "", "post_date": "01/17/2018 17:50:37", "content": "Several things I think are important as starting points: 1) define/guess image magnification; 2) define what parts are cells and what are imaging artefacts; 3) figure out whether the nucleus is stained or not; 4) figure out whether the nucleus is the only thing that's stained, as sometimes whole cells are part of the signal.
", "votes": null, "replies": [] }, { "id": 272318, "author_name": "jenishah", "author_url": "", "post_date": "01/22/2018 20:27:09", "content": "Thank you so much for sharing this information. I was clueless as how to find/interpret the nuclei.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "270043": "When considering my observations, keep in mind that I am not a cell biologist - I only play one on TV.\n\nConsider this when creating image masks:\n\n1) nuclei will be stained in most images, and the rest of cell will not be stained. Most of the time you will see is bunch of nuclei but we should draw an imaginary line for cell membrane that is around the nucleus. In some images that cell membrane line will be visible.\n\n2) Nuclei are not uniformly stained, and the granularity of staining depends both on the type of cell and the exact staining method used. In particular, sometimes there will be 1-2 dots inside a nucleus that are nucleoli (singular is nucleolus). Nucleoli can be unstained while the rest of nucleus is stained, though sometimes the opposite is true.\n\n3) Nuclei of same cells have similar shapes and sizes, though they may seem misshapen at times because of the way cells are laying on the surface. At the same time there are imaging artifacts - sometimes literally specks of dust - that will give strong signal, but are not part of cells. A simple way to exclude those would be to calculate an average area of nuclei and exclude the signal that is smaller than it by some ratio. Also, \"dots\" that have much stronger signal than the rest (5-10x) are almost never cells. Finally, cells also tend to clump together and there should be a way to split three nuclei of tightly packed cells even when there is no obvious space between them.\n\n4) Sometimes it is the rest of cell - the cytoplasm - that is stained rather than the nucleus, or in addition to it. For example, in image cb4df20a83b2f38b394c67f1d9d4aef29f9794d5345da3576318374ec3a11490 the nucleus is a relatively unstained part in the middle rather than strongly stained dots around it.\n\n5) Images have different magnification: some of them are at 20-50x, others ate 200-400x. Needless to say, it is much easier to define nuclei precisely at higher magnification.", "270060": "Several things I think are important as starting points: 1) define/guess image magnification; 2) define what parts are cells and what are imaging artefacts; 3) figure out whether the nucleus is stained or not; 4) figure out whether the nucleus is the only thing that's stained, as sometimes whole cells are part of the signal.", "272318": "Thank you so much for sharing this information. I was clueless as how to find/interpret the nuclei." }, "source": "meta" }