The data tab says:
\n\nsamples are represented by four filters (stored as individual files), the protein of interest (green) plus three cellular landmarks: nucleus (blue), microtubules (red), endoplasmic reticulum (yellow). The green filter should hence be used to predict the label, and the other filters are used as references.
\n\nDoes anyone know what this means? Are the different filters just showing images illuminated by (rgby) light wvelengths or are the colors just labels referring to other image filtering mechanisms? So, are the stated colors physically meaningful?
\n\nAre the green images just showing the protein of interest, by some “microscope” technology? Somehow they get rid of the other junk?
\n\nAre the tiff images 3-channel? If so, does that conflict with the label colors?
\n\nThese are fairly silly questions but I don’t know anything a out the field and I might benefit from a little knowledge. I’ll take a good reference if offered.
", "messages": [ { "id": "414116", "postDate": "11/02/2018 07:04:56", "content": "The data tab says:
\n\nsamples are represented by four filters (stored as individual files), the protein of interest (green) plus three cellular landmarks: nucleus (blue), microtubules (red), endoplasmic reticulum (yellow). The green filter should hence be used to predict the label, and the other filters are used as references.
\n\nDoes anyone know what this means? Are the different filters just showing images illuminated by (rgby) light wvelengths or are the colors just labels referring to other image filtering mechanisms? So, are the stated colors physically meaningful?
\n\nAre the green images just showing the protein of interest, by some “microscope” technology? Somehow they get rid of the other junk?
\n\nAre the tiff images 3-channel? If so, does that conflict with the label colors?
\n\nThese are fairly silly questions but I don’t know anything a out the field and I might benefit from a little knowledge. I’ll take a good reference if offered.
", "rawMarkdown": "The data tab says:\n\nsamples are represented by four filters (stored as individual files), the protein of interest (green) plus three cellular landmarks: nucleus (blue), microtubules (red), endoplasmic reticulum (yellow). The green filter should hence be used to predict the label, and the other filters are used as references.\n\nDoes anyone know what this means? Are the different filters just showing images illuminated by (rgby) light wvelengths or are the colors just labels referring to other image filtering mechanisms? So, are the stated colors physically meaningful?\n\nAre the green images just showing the protein of interest, by some “microscope” technology? Somehow they get rid of the other junk?\n\nAre the tiff images 3-channel? If so, does that conflict with the label colors?\n\nThese are fairly silly questions but I don’t know anything a out the field and I might benefit from a little knowledge. I’ll take a good reference if offered.", "votes": null }, { "id": "414141", "postDate": "11/02/2018 07:41:12", "content": "Blue, red, and yellow are always the same part of the cell in all images. Green is the part of the cell indicated by the labels.
\n\nThe tiff images are 1 channel with _red, _blue, etc.. like the png files.
", "rawMarkdown": "Blue, red, and yellow are always the same part of the cell in all images. Green is the part of the cell indicated by the labels. \n\nThe tiff images are 1 channel with _red, _blue, etc.. like the png files.", "votes": null }, { "id": "414375", "postDate": "11/02/2018 16:12:26", "content": "All images were obtained by microscopy. The only difference is in antibodies used for various cell parts.
\n\nThink of red, blue and yellow images as landmarks. If cells were houses, red would tell you where the roof is, blue would be doors, and yellow would be sides. Now you need to predict where a pollster visiting your house would attach a note, which is what green channel is.
\n\nIt is not a perfect analogy because in cell images red and yellow overlap to a good degree unlike in my example. Nobody would leave a note on the roof either, and one could never leave enough notes to cover the whole door (let alone the whole side), but hopefully the explanation is still useful to you.
", "rawMarkdown": "All images were obtained by microscopy. The only difference is in antibodies used for various cell parts.\n\nThink of red, blue and yellow images as landmarks. If cells were houses, red would tell you where the roof is, blue would be doors, and yellow would be sides. Now you need to predict where a pollster visiting your house would attach a note, which is what green channel is.\n\nIt is not a perfect analogy because in cell images red and yellow overlap to a good degree unlike in my example. Nobody would leave a note on the roof either, and one could never leave enough notes to cover the whole door (let alone the whole side), but hopefully the explanation is still useful to you.", "votes": null }, { "id": "414406", "postDate": "11/02/2018 17:38:22", "content": "Reading the above comments and other stuff, I summarize as follows: \n- red, blue, yellow images are formed by staining with antibodies that bond to 3 common structures in the cells. When bathed in light of a certain wahelength, these antibodies flouresce. These images tell you where the main structures are located - where the cell is. \n- the green images are responding to a stain that attaches to a protein, thus marking out the location of the protein.\n- all 4 images are formed by changing the excitation and measured light filters, while the tissue sample remains in one place, so that the 4 images give a precise measurement of where in the cell the protein is located relative to its main structures.\n- I doubt that the light color actually represents the light wavelength measured - the labels are “false”.\n- the measured light is narrow band - basically a peak in the spectrum, so there is no color information in the images. I guess that it is fine to treat each image as 1 channel, not three.\n- the whole data set deals with just one protein whose identity is not known - perhaps to limit using specific knowledge in the field.
\n\nProbably mostly wrong - I hope someone clarifies if that is the case.
\n\nIt’s a very interesting field that I never really heard of - one of the main reasons I participate in Kaggle
", "rawMarkdown": "Reading the above comments and other stuff, I summarize as follows: \n- red, blue, yellow images are formed by staining with antibodies that bond to 3 common structures in the cells. When bathed in light of a certain wahelength, these antibodies flouresce. These images tell you where the main structures are located - where the cell is. \n- the green images are responding to a stain that attaches to a protein, thus marking out the location of the protein.\n- all 4 images are formed by changing the excitation and measured light filters, while the tissue sample remains in one place, so that the 4 images give a precise measurement of where in the cell the protein is located relative to its main structures.\n- I doubt that the light color actually represents the light wavelength measured - the labels are “false”.\n- the measured light is narrow band - basically a peak in the spectrum, so there is no color information in the images. I guess that it is fine to treat each image as 1 channel, not three.\n- the whole data set deals with just one protein whose identity is not known - perhaps to limit using specific knowledge in the field.\n\nProbably mostly wrong - I hope someone clarifies if that is the case.\n\nIt’s a very interesting field that I never really heard of - one of the main reasons I participate in Kaggle", "votes": null }, { "id": "414525", "postDate": "11/02/2018 23:16:49", "content": "https://www.proteinatlas.org/about/assays+annotation#if
\n\nIn order to facilitate the annotation of the subcellular localization of the protein targeted by the HPA antibody, the cells are also stained with reference markers. The following probes/organelles are used as references; (i) DAPI for the nucleus, (ii) anti-tubulin antibody as internal control and marker of microtubules, and (iii) anti-calreticulin or anti-KDEL for the endoplasmic reticulum (ER).
\n\nThe resulting confocal images are single slice images representing one optical section of the cells. The microscope settings are optimized for each sample. The different organelle probes are displayed as different channels in the multicolor images; the HPA antibody staining is shown in green, nuclear stain in blue, microtubules in red and ER in yellow.
", "rawMarkdown": "https://www.proteinatlas.org/about/assays+annotation#if\n\n\nIn order to facilitate the annotation of the subcellular localization of the protein targeted by the HPA antibody, the cells are also stained with reference markers. The following probes/organelles are used as references; (i) DAPI for the nucleus, (ii) anti-tubulin antibody as internal control and marker of microtubules, and (iii) anti-calreticulin or anti-KDEL for the endoplasmic reticulum (ER).\n\nThe resulting confocal images are single slice images representing one optical section of the cells. The microscope settings are optimized for each sample. The different organelle probes are displayed as different channels in the multicolor images; the HPA antibody staining is shown in green, nuclear stain in blue, microtubules in red and ER in yellow.", "votes": null }, { "id": "414536", "postDate": "11/02/2018 23:33:43", "content": "\n\nFigure 3: Anti-ZYX (HPA004835) shows staining of focal adhesions in A-431 cells. By clicking on the different channels, the images can be visualized with one or several reference markers. A) antibody staining shown in green, B) addition of DAPI nuclear reference in blue, C) addition of microtubules reference in red and D) addition of endoplasmic reticulum (ER) reference in yellow.
\n\n", "rawMarkdown": "see also:\nhttps://www.sigmaaldrich.com/life-science/cell-biology/antibodies/prestige-antibodies/prestige-antibodies-in-immunofluorescence-applications.html\n\nFigure 3: Anti-ZYX (HPA004835) shows staining of focal adhesions in A-431 cells. By clicking on the different channels, the images can be visualized with one or several reference markers. A) antibody staining shown in green, B) addition of DAPI nuclear reference in blue, C) addition of microtubules reference in red and D) addition of endoplasmic reticulum (ER) reference in yellow.\n\n\nhttps://www.sigmaaldrich.com/content/dam/sigma-aldrich/life-science/cell-biology/antibodies/prestige-antibodies/staining-of-focal-adhesions.gif", "votes": null }, { "id": "414642", "postDate": "11/03/2018 09:00:08", "content": "when I read that three colors were “there for reference”, I thought that they might be taken less seriously, or som omitted from training. But in reality they are all important info and all should be used.
", "rawMarkdown": "when I read that three colors were “there for reference”, I thought that they might be taken less seriously, or som omitted from training. But in reality they are all important info and all should be used.", "votes": null } ], "comments": [ { "id": 414141, "author_name": "ldm314", "author_url": "", "post_date": "11/02/2018 07:41:12", "content": "Blue, red, and yellow are always the same part of the cell in all images. Green is the part of the cell indicated by the labels.
\n\nThe tiff images are 1 channel with _red, _blue, etc.. like the png files.
", "votes": null, "replies": [] }, { "id": 414375, "author_name": "tilii7", "author_url": "", "post_date": "11/02/2018 16:12:26", "content": "All images were obtained by microscopy. The only difference is in antibodies used for various cell parts.
\n\nThink of red, blue and yellow images as landmarks. If cells were houses, red would tell you where the roof is, blue would be doors, and yellow would be sides. Now you need to predict where a pollster visiting your house would attach a note, which is what green channel is.
\n\nIt is not a perfect analogy because in cell images red and yellow overlap to a good degree unlike in my example. Nobody would leave a note on the roof either, and one could never leave enough notes to cover the whole door (let alone the whole side), but hopefully the explanation is still useful to you.
", "votes": null, "replies": [] }, { "id": 414406, "author_name": "petewills", "author_url": "", "post_date": "11/02/2018 17:38:22", "content": "Reading the above comments and other stuff, I summarize as follows: \n- red, blue, yellow images are formed by staining with antibodies that bond to 3 common structures in the cells. When bathed in light of a certain wahelength, these antibodies flouresce. These images tell you where the main structures are located - where the cell is. \n- the green images are responding to a stain that attaches to a protein, thus marking out the location of the protein.\n- all 4 images are formed by changing the excitation and measured light filters, while the tissue sample remains in one place, so that the 4 images give a precise measurement of where in the cell the protein is located relative to its main structures.\n- I doubt that the light color actually represents the light wavelength measured - the labels are “false”.\n- the measured light is narrow band - basically a peak in the spectrum, so there is no color information in the images. I guess that it is fine to treat each image as 1 channel, not three.\n- the whole data set deals with just one protein whose identity is not known - perhaps to limit using specific knowledge in the field.
\n\nProbably mostly wrong - I hope someone clarifies if that is the case.
\n\nIt’s a very interesting field that I never really heard of - one of the main reasons I participate in Kaggle
", "votes": null, "replies": [] }, { "id": 414525, "author_name": "hengck23", "author_url": "", "post_date": "11/02/2018 23:16:49", "content": "https://www.proteinatlas.org/about/assays+annotation#if
\n\nIn order to facilitate the annotation of the subcellular localization of the protein targeted by the HPA antibody, the cells are also stained with reference markers. The following probes/organelles are used as references; (i) DAPI for the nucleus, (ii) anti-tubulin antibody as internal control and marker of microtubules, and (iii) anti-calreticulin or anti-KDEL for the endoplasmic reticulum (ER).
\n\nThe resulting confocal images are single slice images representing one optical section of the cells. The microscope settings are optimized for each sample. The different organelle probes are displayed as different channels in the multicolor images; the HPA antibody staining is shown in green, nuclear stain in blue, microtubules in red and ER in yellow.
", "votes": null, "replies": [] }, { "id": 414536, "author_name": "hengck23", "author_url": "", "post_date": "11/02/2018 23:33:43", "content": "\n\nFigure 3: Anti-ZYX (HPA004835) shows staining of focal adhesions in A-431 cells. By clicking on the different channels, the images can be visualized with one or several reference markers. A) antibody staining shown in green, B) addition of DAPI nuclear reference in blue, C) addition of microtubules reference in red and D) addition of endoplasmic reticulum (ER) reference in yellow.
\n\n", "votes": null, "replies": [] }, { "id": 414642, "author_name": "petewills", "author_url": "", "post_date": "11/03/2018 09:00:08", "content": "when I read that three colors were “there for reference”, I thought that they might be taken less seriously, or som omitted from training. But in reality they are all important info and all should be used.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "414116": "The data tab says:\n\nsamples are represented by four filters (stored as individual files), the protein of interest (green) plus three cellular landmarks: nucleus (blue), microtubules (red), endoplasmic reticulum (yellow). The green filter should hence be used to predict the label, and the other filters are used as references.\n\nDoes anyone know what this means? Are the different filters just showing images illuminated by (rgby) light wvelengths or are the colors just labels referring to other image filtering mechanisms? So, are the stated colors physically meaningful?\n\nAre the green images just showing the protein of interest, by some “microscope” technology? Somehow they get rid of the other junk?\n\nAre the tiff images 3-channel? If so, does that conflict with the label colors?\n\nThese are fairly silly questions but I don’t know anything a out the field and I might benefit from a little knowledge. I’ll take a good reference if offered.", "414141": "Blue, red, and yellow are always the same part of the cell in all images. Green is the part of the cell indicated by the labels. \n\nThe tiff images are 1 channel with _red, _blue, etc.. like the png files.", "414375": "All images were obtained by microscopy. The only difference is in antibodies used for various cell parts.\n\nThink of red, blue and yellow images as landmarks. If cells were houses, red would tell you where the roof is, blue would be doors, and yellow would be sides. Now you need to predict where a pollster visiting your house would attach a note, which is what green channel is.\n\nIt is not a perfect analogy because in cell images red and yellow overlap to a good degree unlike in my example. Nobody would leave a note on the roof either, and one could never leave enough notes to cover the whole door (let alone the whole side), but hopefully the explanation is still useful to you.", "414406": "Reading the above comments and other stuff, I summarize as follows: \n- red, blue, yellow images are formed by staining with antibodies that bond to 3 common structures in the cells. When bathed in light of a certain wahelength, these antibodies flouresce. These images tell you where the main structures are located - where the cell is. \n- the green images are responding to a stain that attaches to a protein, thus marking out the location of the protein.\n- all 4 images are formed by changing the excitation and measured light filters, while the tissue sample remains in one place, so that the 4 images give a precise measurement of where in the cell the protein is located relative to its main structures.\n- I doubt that the light color actually represents the light wavelength measured - the labels are “false”.\n- the measured light is narrow band - basically a peak in the spectrum, so there is no color information in the images. I guess that it is fine to treat each image as 1 channel, not three.\n- the whole data set deals with just one protein whose identity is not known - perhaps to limit using specific knowledge in the field.\n\nProbably mostly wrong - I hope someone clarifies if that is the case.\n\nIt’s a very interesting field that I never really heard of - one of the main reasons I participate in Kaggle", "414525": "https://www.proteinatlas.org/about/assays+annotation#if\n\n\nIn order to facilitate the annotation of the subcellular localization of the protein targeted by the HPA antibody, the cells are also stained with reference markers. The following probes/organelles are used as references; (i) DAPI for the nucleus, (ii) anti-tubulin antibody as internal control and marker of microtubules, and (iii) anti-calreticulin or anti-KDEL for the endoplasmic reticulum (ER).\n\nThe resulting confocal images are single slice images representing one optical section of the cells. The microscope settings are optimized for each sample. The different organelle probes are displayed as different channels in the multicolor images; the HPA antibody staining is shown in green, nuclear stain in blue, microtubules in red and ER in yellow.", "414536": "see also:\nhttps://www.sigmaaldrich.com/life-science/cell-biology/antibodies/prestige-antibodies/prestige-antibodies-in-immunofluorescence-applications.html\n\nFigure 3: Anti-ZYX (HPA004835) shows staining of focal adhesions in A-431 cells. By clicking on the different channels, the images can be visualized with one or several reference markers. A) antibody staining shown in green, B) addition of DAPI nuclear reference in blue, C) addition of microtubules reference in red and D) addition of endoplasmic reticulum (ER) reference in yellow.\n\n\nhttps://www.sigmaaldrich.com/content/dam/sigma-aldrich/life-science/cell-biology/antibodies/prestige-antibodies/staining-of-focal-adhesions.gif", "414642": "when I read that three colors were “there for reference”, I thought that they might be taken less seriously, or som omitted from training. But in reality they are all important info and all should be used." }, "source": "meta" }![https://www.sigmaaldrich.com/content/dam/sigma-aldrich/life-science/cell-biology/antibodies/prestige-antibodies/staining-of-focal-adhesions.gif](\"https://www.sigmaaldrich.com/content/dam/sigma-aldrich/life-science/cell-biology/antibodies/prestige-antibodies/staining-of-focal-adhesions.gif\"){kind=link}