46:I[191445,["/_next/static/immutable/chunks/1gnv3tznf_evh.js","/_next/static/immutable/chunks/1t2n4mlwbtz4z.js","/_next/static/immutable/chunks/0v6-3q16afmwn.js","/_next/static/immutable/chunks/1t21y9i8gfouc.js","/_next/static/immutable/chunks/13wylyv8p_3ah.js","/_next/static/immutable/chunks/3dbi90ge1lq2d.js","/_next/static/immutable/chunks/09_zswn44hdxd.js","/_next/static/immutable/chunks/2ri0u_hhdt48v.js","/_next/static/immutable/chunks/3rsnyfmzi6oij.js","/_next/static/immutable/chunks/0um9a0fy8puzs.js","/_next/static/immutable/chunks/28z-cpcwnf3jk.js","/_next/static/immutable/chunks/1qtf0679ftdt9.js","/_next/static/immutable/chunks/39rtjowexcmc2.js","/_next/static/immutable/chunks/2j--adssq6-u5.js","/_next/static/immutable/chunks/3vj_x3co-p-xb.js","/_next/static/immutable/chunks/0tfe231gghziz.js","/_next/static/immutable/chunks/2xqpa5c7ogclt.js","/_next/static/immutable/chunks/2x03jtn4ys_wx.js","/_next/static/immutable/chunks/3ljsqu27kfk5s.js","/_next/static/immutable/chunks/306xyefo9piez.js","/_next/static/immutable/chunks/1e0jyw4aho8bn.js","/_next/static/immutable/chunks/2b7liv95ur9vy.js","/_next/static/immutable/chunks/3qwukaf1anwoc.js","/_next/static/immutable/chunks/2r69nv9dqp2-d.js","/_next/static/immutable/chunks/0n4ud2tq5ra38.js","/_next/static/immutable/chunks/28eg3-lj7mp4r.js","/_next/static/immutable/chunks/2lfxvp7uhhdj1.js","/_next/static/immutable/chunks/2nmsuluddca8y.js","/_next/static/immutable/chunks/31m6mwk8-hjej.js","/_next/static/immutable/chunks/42gxo7rog0ln2.js","/_next/static/immutable/chunks/2xrpvk2x3o4ry.js","/_next/static/immutable/chunks/0lg0h28s3z3jk.js","/_next/static/immutable/chunks/3a7he7r-443qg.js","/_next/static/immutable/chunks/0by0nrve34sen.js","/_next/static/immutable/chunks/1qoagqotwft5_.js","/_next/static/immutable/chunks/1q7kned6j2qte.js","/_next/static/immutable/chunks/3jho5pc1xhg-w.js","/_next/static/immutable/chunks/3-y1wg0q2kip1.js","/_next/static/immutable/chunks/40q3jrojjt2a4.js","/_next/static/immutable/chunks/1nv14oq84mc1d.js"],"WorkTabProvider"] 47:I[532546,["/_next/static/immutable/chunks/1gnv3tznf_evh.js","/_next/static/immutable/chunks/1t2n4mlwbtz4z.js","/_next/static/immutable/chunks/0v6-3q16afmwn.js","/_next/static/immutable/chunks/1t21y9i8gfouc.js","/_next/static/immutable/chunks/13wylyv8p_3ah.js","/_next/static/immutable/chunks/3dbi90ge1lq2d.js","/_next/static/immutable/chunks/09_zswn44hdxd.js","/_next/static/immutable/chunks/2ri0u_hhdt48v.js","/_next/static/immutable/chunks/3rsnyfmzi6oij.js","/_next/static/immutable/chunks/0um9a0fy8puzs.js","/_next/static/immutable/chunks/28z-cpcwnf3jk.js","/_next/static/immutable/chunks/1qtf0679ftdt9.js","/_next/static/immutable/chunks/39rtjowexcmc2.js","/_next/static/immutable/chunks/2j--adssq6-u5.js","/_next/static/immutable/chunks/3vj_x3co-p-xb.js","/_next/static/immutable/chunks/0tfe231gghziz.js","/_next/static/immutable/chunks/2xqpa5c7ogclt.js","/_next/static/immutable/chunks/2x03jtn4ys_wx.js","/_next/static/immutable/chunks/3ljsqu27kfk5s.js","/_next/static/immutable/chunks/306xyefo9piez.js","/_next/static/immutable/chunks/1e0jyw4aho8bn.js","/_next/static/immutable/chunks/2b7liv95ur9vy.js","/_next/static/immutable/chunks/3qwukaf1anwoc.js","/_next/static/immutable/chunks/2r69nv9dqp2-d.js","/_next/static/immutable/chunks/0n4ud2tq5ra38.js","/_next/static/immutable/chunks/28eg3-lj7mp4r.js","/_next/static/immutable/chunks/2lfxvp7uhhdj1.js","/_next/static/immutable/chunks/2nmsuluddca8y.js","/_next/static/immutable/chunks/31m6mwk8-hjej.js","/_next/static/immutable/chunks/42gxo7rog0ln2.js","/_next/static/immutable/chunks/2xrpvk2x3o4ry.js","/_next/static/immutable/chunks/0lg0h28s3z3jk.js","/_next/static/immutable/chunks/3a7he7r-443qg.js","/_next/static/immutable/chunks/0by0nrve34sen.js","/_next/static/immutable/chunks/1qoagqotwft5_.js","/_next/static/immutable/chunks/1q7kned6j2qte.js","/_next/static/immutable/chunks/3jho5pc1xhg-w.js","/_next/static/immutable/chunks/3-y1wg0q2kip1.js","/_next/static/immutable/chunks/40q3jrojjt2a4.js","/_next/static/immutable/chunks/1nv14oq84mc1d.js"],"WorkHeaderProposal"] 53:I[127600,["/_next/static/immutable/chunks/1gnv3tznf_evh.js","/_next/static/immutable/chunks/1t2n4mlwbtz4z.js","/_next/static/immutable/chunks/0v6-3q16afmwn.js","/_next/static/immutable/chunks/1t21y9i8gfouc.js","/_next/static/immutable/chunks/13wylyv8p_3ah.js","/_next/static/immutable/chunks/3dbi90ge1lq2d.js","/_next/static/immutable/chunks/09_zswn44hdxd.js","/_next/static/immutable/chunks/2ri0u_hhdt48v.js","/_next/static/immutable/chunks/3rsnyfmzi6oij.js","/_next/static/immutable/chunks/0um9a0fy8puzs.js","/_next/static/immutable/chunks/28z-cpcwnf3jk.js","/_next/static/immutable/chunks/1qtf0679ftdt9.js","/_next/static/immutable/chunks/39rtjowexcmc2.js","/_next/static/immutable/chunks/2j--adssq6-u5.js","/_next/static/immutable/chunks/3vj_x3co-p-xb.js","/_next/static/immutable/chunks/0tfe231gghziz.js","/_next/static/immutable/chunks/2xqpa5c7ogclt.js","/_next/static/immutable/chunks/2x03jtn4ys_wx.js","/_next/static/immutable/chunks/3ljsqu27kfk5s.js","/_next/static/immutable/chunks/306xyefo9piez.js","/_next/static/immutable/chunks/1e0jyw4aho8bn.js","/_next/static/immutable/chunks/2b7liv95ur9vy.js","/_next/static/immutable/chunks/3qwukaf1anwoc.js","/_next/static/immutable/chunks/2r69nv9dqp2-d.js","/_next/static/immutable/chunks/0n4ud2tq5ra38.js","/_next/static/immutable/chunks/28eg3-lj7mp4r.js","/_next/static/immutable/chunks/2lfxvp7uhhdj1.js","/_next/static/immutable/chunks/2nmsuluddca8y.js","/_next/static/immutable/chunks/31m6mwk8-hjej.js","/_next/static/immutable/chunks/42gxo7rog0ln2.js","/_next/static/immutable/chunks/2xrpvk2x3o4ry.js","/_next/static/immutable/chunks/0lg0h28s3z3jk.js","/_next/static/immutable/chunks/3a7he7r-443qg.js","/_next/static/immutable/chunks/0by0nrve34sen.js","/_next/static/immutable/chunks/1qoagqotwft5_.js","/_next/static/immutable/chunks/1q7kned6j2qte.js","/_next/static/immutable/chunks/3jho5pc1xhg-w.js","/_next/static/immutable/chunks/3-y1wg0q2kip1.js","/_next/static/immutable/chunks/40q3jrojjt2a4.js","/_next/static/immutable/chunks/1nv14oq84mc1d.js","/_next/static/immutable/chunks/25pn5-4lvzr3y.js","/_next/static/immutable/chunks/3a_j64ifthr1h.js","/_next/static/immutable/chunks/23ycn2rhsarmy.js","/_next/static/immutable/chunks/2lkwbg84bg5zx.js","/_next/static/immutable/chunks/2umzq3lq261-2.js","/_next/static/immutable/chunks/022pez2xgib6m.js","/_next/static/immutable/chunks/2hxxo2gyb459l.js","/_next/static/immutable/chunks/2_v15tiz7-v2_.js","/_next/static/immutable/chunks/21658sh0k0q42.js","/_next/static/immutable/chunks/13tnhw8p7524f.js","/_next/static/immutable/chunks/0kuzg7x1feq0o.js","/_next/static/immutable/chunks/19zgw2d_e3tu1.js","/_next/static/immutable/chunks/2ddrbcod4yybp.js"],"FundDocument"] 48:T11fa,Early Automated Detection of Cell Culture Failure from Time-Lapse Imaging. 1. Abstract This study will consists of an incubated mammalian cell type over their weekly growth period. The purpose of the study is to determine whether ML models can determine whether cell growth will fail or succeed in a rapid timeline. The work will be carried out by Andre Childs (PI) at the Dallas Maker Space’s cell culture laboratory. The duration of the project will be last 12 weeks from funding to delivery. Raw images with annotations, protocols and processing logs will be licensed to the CUI Fund Pilot under written agreement and archived openly. 2. Introduction Cell cultures fail through overgrowth, poor attachment, contamination and environmental and media changes. These failure are currently found after inspection from human intervention days after a culture begins or when media changes are required. [1] This is a large loss to studies as it wastes cultured cells, media, consumable but most of all the scientists time as weekends are often required to inspect cells at both the academic and industrial level. With Hypotheses H1 . Can ML algorithms determine cell responses before human intervention is required. H2. Can a model trained from the first 24 hours predict which well have failed before 96 hours have passed or become visible to assay developmental scientists. 3. Methods Participants/Sample The target sample will be HEK293 Mammalian cells. Cell growth wells contaminated, failed and successful will provided. Cell lines will be purchased from the ATTC and studied in within 4 different independent experiments seeded on different days. The experiment will consist of 1. Healthy seeded at recommended density given the low doses of nutrients. 2. Healthy cells seeded at the recommended density and the proper amount of nutrients. 3. Healthy Cells given higher doses of nutrients. 4. Healthy cells seeded at the recommended density and proper amount of nutrients spiked with microbial contamination. Temperatures will remain consistent throughout each of the different experimental procedures. lmaging will occur with a bottom up microscope at given times of 1,2,3, 5 and 7 days. Incubator lapse image videos will also be provided to the licensed CUI Fund Pilot. Materials and Procedures Cell culture. Cells grow in DMEM (high glucose) with 10% fetal bovine serum at 37 °C and 5% CO2 and are passaged at 70–90% confluence with trypsin-EDTA. Antibiotics are omitted from one passage before seeding onward, so that contamination is not masked. Medium and serum lot numbers are recorded. Plate setup. Healthy wells are seeded at different densities as specified in the Sample section density. Condition positions follow a pre-generated randomized layout. Plates wells use a breathable sealing film and sit in a closed secondary container in the incubator; they are autoclaved after the final image. ( Or left to continue growing until a failure mode occurs to increase the amount of cellular data.) Imaging. All images are taken on an inverted phase-contrast microscope. Exposure, gain and light intensity will be fixed. No plate will spend more than 10 minutes out of the incubator per session. The time out of the incubator will be logged. All images will be taken with the microscopy fields and 5 images will be taken per well. End-point measurements. At 96 h, the plate is photographed against a color reference card to record medium color. Every well is then harvested with its floating cells and counted with trypan blue on a hemocytometer, in duplicate. Annotations for the dataset. Every image will be listed with the experiment and plate, condition, time, passage. Planned Analyses In our experimental pipeline analysis the raw images will be provided archived and unchanged. Two Algorithms will be built and compared. One will consist of an anomaly detector, learning the cells growth curve and spread of the features. Ethics and Data Management This study uses an established and commercially available distributed human cell line and involved no patients, patient identifiable data or animals. Data management will consist of seperate raw images, kept under lock and key. Admin privileges are only allowed to view the working drive. 5. Budget Provide an itemized list of anticipated costs (materials, personnel, etc.). These are not binding and are able to be adjusted based on the results of the pilot data. 6. References List all references used in the manuscript using Nature style (e.g. numbered in-text citations and alphabetical references at the end of the manuscript).49:T1220,Early Automated Detection of Cell Culture Failure from Time-Lapse Imaging. 1. Abstract This study will consists of an incubated mammalian cell type over their weekly growth period. The purpose of the study is to determine whether ML models can determine whether cell growth will fail or succeed in a rapid timeline. The work will be carried out by Andre Childs (PI) at the Dallas Maker Space’s cell culture laboratory. The duration of the project will be last 12 weeks from funding to delivery. Raw images with annotations, protocols and processing logs will be licensed to the CUI Fund Pilot under written agreement and archived openly. 2. Introduction Cell cultures fail through overgrowth, poor attachment, contamination and environmental and media changes. These failure are currently found after inspection from human intervention days after a culture begins or when media changes are required. [1] This is a large loss to studies as it wastes cultured cells, media, consumable but most of all the scientists time as weekends are often required to inspect cells at both the academic and industrial level. With Hypotheses H1 . Can ML algorithms determine cell responses before human intervention is required. H2. Can a model trained from the first 24 hours predict which well have failed before 96 hours have passed or become visible to assay developmental scientists. 3. Methods Participants/Sample The target sample will be HEK293 Mammalian cells. Cell growth wells contaminated, failed and successful will provided. Cell lines will be purchased from the ATTC and studied in within 4 different independent experiments seeded on different days. The experiment will consist of 1. Healthy seeded at recommended density given the low doses of nutrients. 2. Healthy cells seeded at the recommended density and the proper amount of nutrients. 3. Healthy Cells given higher doses of nutrients. 4. Healthy cells seeded at the recommended density and proper amount of nutrients spiked with microbial contamination. Temperatures will remain consistent throughout each of the different experimental procedures. lmaging will occur with a bottom up microscope at given times of 1,2,3, 5 and 7 days. Incubator lapse image videos will also be provided to the licensed CUI Fund Pilot. Materials and Procedures Cell culture. Cells grow in DMEM (high glucose) with 10% fetal bovine serum at 37 °C and 5% CO2 and are passaged at 70–90% confluence with trypsin-EDTA. Antibiotics are omitted from one passage before seeding onward, so that contamination is not masked. Medium and serum lot numbers are recorded. Plate setup. Healthy wells are seeded at different densities as specified in the Sample section density. Condition positions follow a pre-generated randomized layout. Plates wells use a breathable sealing film and sit in a closed secondary container in the incubator; they are autoclaved after the final image. ( Or left to continue growing until a failure mode occurs to increase the amount of cellular data.) Imaging. All images are taken on an inverted phase-contrast microscope. Exposure, gain and light intensity will be fixed. No plate will spend more than 10 minutes out of the incubator per session. The time out of the incubator will be logged. All images will be taken with the microscopy fields and 5 images will be taken per well. End-point measurements. At 96 h, the plate is photographed against a color reference card to record medium color. Every well is then harvested with its floating cells and counted with trypan blue on a hemocytometer, in duplicate. Annotations for the dataset. Every image will be listed with the experiment and plate, condition, time, passage. Planned Analyses In our experimental pipeline analysis the raw images will be provided archived and unchanged. Two Algorithms will be built and compared. One will consist of an anomaly detector, learning the cells growth curve and spread of the features. Ethics and Data Management This study uses an established and commercially available distributed human cell line and involved no patients, patient identifiable data or animals. Data management will consist of seperate raw images, kept under lock and key. Admin privileges are only allowed to view the working drive. 5. Budget Provide an itemized list of anticipated costs (materials, personnel, etc.). These are not binding and are able to be adjusted based on the results of the pilot data. 6. References List all references used in the manuscript using Nature style (e.g. numbered in-text citations and alphabetical references at the end of the manuscript).4a:T22f9,{"type":"doc","content":[{"type":"heading","attrs":{"id":"22925bb2-8f35-4a4d-9f7a-32798463bbad","data-toc-id":"22925bb2-8f35-4a4d-9f7a-32798463bbad","textAlign":"left","level":1},"content":[{"type":"text","text":"Early Automated Detection of Cell Culture Failure from Time-Lapse Imaging. "}]},{"type":"heading","attrs":{"id":"102280b4-4c98-45b4-9bc3-29b415268885","data-toc-id":"102280b4-4c98-45b4-9bc3-29b415268885","textAlign":"left","level":1},"content":[{"type":"text","text":"1. Abstract"}]},{"type":"paragraph","attrs":{"id":"3e85ceae-55f1-43db-b440-a80df2f693e9","class":null,"textAlign":"left"},"content":[{"type":"hardBreak"},{"type":"text","text":"This study will consists of an incubated mammalian cell type over their weekly growth period. The purpose of the study is to determine whether ML models can determine whether cell growth will fail or succeed in a rapid timeline. The work will be carried out by Andre Childs (PI) at the Dallas Maker Space’s cell culture laboratory. The duration of the project will be last 12 weeks from funding to delivery. Raw images with annotations, protocols and processing logs will be licensed to the CUI Fund Pilot under written agreement and archived openly. "}]},{"type":"heading","attrs":{"id":"a75a40d2-a3e5-458a-9094-1a137760f013","data-toc-id":"a75a40d2-a3e5-458a-9094-1a137760f013","textAlign":null,"level":1},"content":[{"type":"text","text":"2. Introduction"}]},{"type":"paragraph","attrs":{"id":"eab0601e-2596-41b2-958d-227927f52356","class":null,"textAlign":"left"},"content":[{"type":"text","text":"Cell cultures fail through overgrowth, poor attachment, contamination and environmental and media changes. These failure are currently found after inspection from human intervention days after a culture begins or when media changes are required. [1] This is a large loss to studies as it wastes cultured cells, media, consumable but most of all the scientists time as weekends are often required to inspect cells at both the academic and industrial level. With "}]},{"type":"heading","attrs":{"id":"1504af7b-fd4b-47b2-96fe-7947c6b08e82","data-toc-id":"1504af7b-fd4b-47b2-96fe-7947c6b08e82","textAlign":"left","level":3},"content":[{"type":"text","text":"Hypotheses"}]},{"type":"paragraph","attrs":{"id":"69cbe293-7114-4615-8090-415936addfbe","class":null,"textAlign":"left"},"content":[{"type":"hardBreak"},{"type":"text","text":"H1 . Can ML algorithms determine cell responses before human intervention is required. "}]},{"type":"paragraph","attrs":{"id":"f13d3a01-e378-4b8b-b2b1-1c4a831b8751","class":null,"textAlign":"left"},"content":[{"type":"text","text":"H2. Can a model trained from the first 24 hours predict which well have failed before 96 hours have passed or become visible to assay developmental scientists."}]},{"type":"heading","attrs":{"id":"5db21052-1080-4fc0-bce1-4b1a0690b84d","data-toc-id":"5db21052-1080-4fc0-bce1-4b1a0690b84d","textAlign":"left","level":2},"content":[{"type":"text","text":"3. Methods"}]},{"type":"heading","attrs":{"id":"39c552e4-4ee6-4f4f-a880-2b998198760a","data-toc-id":"39c552e4-4ee6-4f4f-a880-2b998198760a","textAlign":"left","level":3},"content":[{"type":"text","text":"Participants/Sample"}]},{"type":"paragraph","attrs":{"id":"e290afab-d7f3-4378-85a5-d382071a1f9a","class":null,"textAlign":"justify"},"content":[{"type":"hardBreak"},{"type":"text","text":"The target sample will be HEK293 Mammalian cells. Cell growth wells contaminated, failed and successful will provided. Cell lines will be purchased from the ATTC and studied in within 4 different independent experiments seeded on different days. The experiment will consist of 1. Healthy seeded at recommended density given the low doses of nutrients. 2. Healthy cells seeded at the recommended density and the proper amount of nutrients. 3. Healthy Cells given higher doses of nutrients. 4. Healthy cells seeded at the recommended density and proper amount of nutrients spiked with microbial contamination. "},{"type":"hardBreak"},{"type":"text","text":"Temperatures will remain consistent throughout each of the different experimental procedures. lmaging will occur with a bottom up microscope at given times of 1,2,3, 5 and 7 days. Incubator lapse image videos will also be provided to the licensed CUI Fund Pilot. "}]},{"type":"heading","attrs":{"id":"62b8f3d6-c707-4b23-8272-1826e8d107db","data-toc-id":"62b8f3d6-c707-4b23-8272-1826e8d107db","textAlign":"left","level":3},"content":[{"type":"text","text":"Materials and Procedures"}]},{"type":"paragraph","attrs":{"id":"8a47998c-8c8b-41e4-8d46-af85adfa9109","class":null,"textAlign":"left"},"content":[{"type":"hardBreak"},{"type":"text","marks":[{"type":"bold"}],"text":"Cell culture."},{"type":"text","text":" Cells grow in DMEM (high glucose) with 10% fetal bovine serum at 37 °C and 5% CO2 and are passaged at 70–90% confluence with trypsin-EDTA. Antibiotics are omitted from one passage before seeding onward, so that contamination is not masked. Medium and serum lot numbers are recorded."},{"type":"hardBreak"},{"type":"text","marks":[{"type":"bold"}],"text":"Plate setup."},{"type":"text","text":" Healthy wells are seeded at different densities as specified in the Sample section density. Condition positions follow a pre-generated randomized layout. Plates wells use a breathable sealing film and sit in a closed secondary container in the incubator; they are autoclaved after the final image. ( Or left to continue growing until a failure mode occurs to increase the amount of cellular data.)"},{"type":"hardBreak"},{"type":"text","marks":[{"type":"bold"}],"text":"Imaging."},{"type":"text","text":" All images are taken on an inverted phase-contrast microscope. Exposure, gain and light intensity will be fixed. No plate will spend more than 10 minutes out of the incubator per session. The time out of the incubator will be logged. All images will be taken with the microscopy fields and 5 images will be taken per well."}]},{"type":"paragraph","attrs":{"id":"f028706d-fc57-49f0-b5aa-bc84e2308233","class":null,"textAlign":"left"},"content":[{"type":"text","marks":[{"type":"bold"}],"text":"End-point measurements."},{"type":"text","text":" At 96 h, the plate is photographed against a color reference card to record medium color. Every well is then harvested with its floating cells and counted with trypan blue on a hemocytometer, in duplicate."},{"type":"hardBreak"},{"type":"text","marks":[{"type":"bold"}],"text":"Annotations for the dataset. "},{"type":"text","text":"Every image will be listed with the experiment and plate, condition, time, passage. "},{"type":"hardBreak"}]},{"type":"heading","attrs":{"id":"5441ad88-0fa0-4c6d-94a3-63a810a56bbd","data-toc-id":"5441ad88-0fa0-4c6d-94a3-63a810a56bbd","textAlign":"left","level":3},"content":[{"type":"text","text":"Planned Analyses"}]},{"type":"paragraph","attrs":{"id":"a8174e24-da95-4fd2-8514-a83f2773a60d","class":null,"textAlign":"left"},"content":[{"type":"hardBreak"},{"type":"text","text":"In our experimental pipeline analysis the raw images will be provided archived and unchanged. Two Algorithms will be built and compared. One will consist of an anomaly detector, learning the cells growth curve and spread of the features. "}]},{"type":"heading","attrs":{"id":"16fe8d55-579c-442d-b1fe-4daddd0cdabc","data-toc-id":"16fe8d55-579c-442d-b1fe-4daddd0cdabc","textAlign":"left","level":3},"content":[{"type":"text","text":"Ethics and Data Management"}]},{"type":"paragraph","attrs":{"id":"90a7cdea-14b2-4857-8704-e5e32d43dae3","class":null,"textAlign":"left"},"content":[{"type":"text","text":"This study uses an established and commercially available distributed human cell line and involved no patients, patient identifiable data or animals. "},{"type":"hardBreak"},{"type":"text","text":"Data management will consist of seperate raw images, kept under lock and key. Admin privileges are only allowed to view the working drive. "}]},{"type":"heading","attrs":{"id":"e32a1c94-5782-4da9-afe1-315a939fe81e","data-toc-id":"e32a1c94-5782-4da9-afe1-315a939fe81e","textAlign":"left","level":2},"content":[{"type":"text","text":"5. Budget"}]},{"type":"paragraph","attrs":{"id":"c835f7ed-c2d5-4a48-b857-e951a9188aac","class":null,"textAlign":"left"},"content":[{"type":"text","text":"Provide an itemized list of anticipated costs (materials, personnel, etc.). These are not binding and are able to be adjusted based on the results of the pilot data."}]},{"type":"heading","attrs":{"id":"d9c90060-9401-49ae-af24-d2ec4c002f31","data-toc-id":"d9c90060-9401-49ae-af24-d2ec4c002f31","textAlign":"left","level":2},"content":[{"type":"text","text":"6. References"}]},{"type":"paragraph","attrs":{"id":"21c387e9-720d-45f2-92f0-fd75d7d84e0b","class":null,"textAlign":"left"},"content":[{"type":"text","text":"List all references used in the manuscript using Nature style (e.g. numbered in-text citations and alphabetical references at the end of the manuscript)."}]}]}4b:T1220,Early Automated Detection of Cell Culture Failure from Time-Lapse Imaging. 1. Abstract This study will consists of an incubated mammalian cell type over their weekly growth period. The purpose of the study is to determine whether ML models can determine whether cell growth will fail or succeed in a rapid timeline. The work will be carried out by Andre Childs (PI) at the Dallas Maker Space’s cell culture laboratory. The duration of the project will be last 12 weeks from funding to delivery. Raw images with annotations, protocols and processing logs will be licensed to the CUI Fund Pilot under written agreement and archived openly. 2. Introduction Cell cultures fail through overgrowth, poor attachment, contamination and environmental and media changes. These failure are currently found after inspection from human intervention days after a culture begins or when media changes are required. [1] This is a large loss to studies as it wastes cultured cells, media, consumable but most of all the scientists time as weekends are often required to inspect cells at both the academic and industrial level. With Hypotheses H1 . Can ML algorithms determine cell responses before human intervention is required. H2. Can a model trained from the first 24 hours predict which well have failed before 96 hours have passed or become visible to assay developmental scientists. 3. Methods Participants/Sample The target sample will be HEK293 Mammalian cells. Cell growth wells contaminated, failed and successful will provided. Cell lines will be purchased from the ATTC and studied in within 4 different independent experiments seeded on different days. The experiment will consist of 1. Healthy seeded at recommended density given the low doses of nutrients. 2. Healthy cells seeded at the recommended density and the proper amount of nutrients. 3. Healthy Cells given higher doses of nutrients. 4. Healthy cells seeded at the recommended density and proper amount of nutrients spiked with microbial contamination. Temperatures will remain consistent throughout each of the different experimental procedures. lmaging will occur with a bottom up microscope at given times of 1,2,3, 5 and 7 days. Incubator lapse image videos will also be provided to the licensed CUI Fund Pilot. Materials and Procedures Cell culture. Cells grow in DMEM (high glucose) with 10% fetal bovine serum at 37 °C and 5% CO2 and are passaged at 70–90% confluence with trypsin-EDTA. Antibiotics are omitted from one passage before seeding onward, so that contamination is not masked. Medium and serum lot numbers are recorded. Plate setup. Healthy wells are seeded at different densities as specified in the Sample section density. Condition positions follow a pre-generated randomized layout. Plates wells use a breathable sealing film and sit in a closed secondary container in the incubator; they are autoclaved after the final image. ( Or left to continue growing until a failure mode occurs to increase the amount of cellular data.) Imaging. All images are taken on an inverted phase-contrast microscope. Exposure, gain and light intensity will be fixed. No plate will spend more than 10 minutes out of the incubator per session. The time out of the incubator will be logged. All images will be taken with the microscopy fields and 5 images will be taken per well. End-point measurements. At 96 h, the plate is photographed against a color reference card to record medium color. Every well is then harvested with its floating cells and counted with trypan blue on a hemocytometer, in duplicate. Annotations for the dataset. Every image will be listed with the experiment and plate, condition, time, passage. Planned Analyses In our experimental pipeline analysis the raw images will be provided archived and unchanged. Two Algorithms will be built and compared. One will consist of an anomaly detector, learning the cells growth curve and spread of the features. Ethics and Data Management This study uses an established and commercially available distributed human cell line and involved no patients, patient identifiable data or animals. Data management will consist of seperate raw images, kept under lock and key. Admin privileges are only allowed to view the working drive. 5. Budget Provide an itemized list of anticipated costs (materials, personnel, etc.). These are not binding and are able to be adjusted based on the results of the pilot data. 6. References List all references used in the manuscript using Nature style (e.g. numbered in-text citations and alphabetical references at the end of the manuscript).4c:T22f9,{"type":"doc","content":[{"type":"heading","attrs":{"id":"22925bb2-8f35-4a4d-9f7a-32798463bbad","data-toc-id":"22925bb2-8f35-4a4d-9f7a-32798463bbad","textAlign":"left","level":1},"content":[{"type":"text","text":"Early Automated Detection of Cell Culture Failure from Time-Lapse Imaging. "}]},{"type":"heading","attrs":{"id":"102280b4-4c98-45b4-9bc3-29b415268885","data-toc-id":"102280b4-4c98-45b4-9bc3-29b415268885","textAlign":"left","level":1},"content":[{"type":"text","text":"1. Abstract"}]},{"type":"paragraph","attrs":{"id":"3e85ceae-55f1-43db-b440-a80df2f693e9","class":null,"textAlign":"left"},"content":[{"type":"hardBreak"},{"type":"text","text":"This study will consists of an incubated mammalian cell type over their weekly growth period. The purpose of the study is to determine whether ML models can determine whether cell growth will fail or succeed in a rapid timeline. The work will be carried out by Andre Childs (PI) at the Dallas Maker Space’s cell culture laboratory. The duration of the project will be last 12 weeks from funding to delivery. Raw images with annotations, protocols and processing logs will be licensed to the CUI Fund Pilot under written agreement and archived openly. "}]},{"type":"heading","attrs":{"id":"a75a40d2-a3e5-458a-9094-1a137760f013","data-toc-id":"a75a40d2-a3e5-458a-9094-1a137760f013","textAlign":null,"level":1},"content":[{"type":"text","text":"2. Introduction"}]},{"type":"paragraph","attrs":{"id":"eab0601e-2596-41b2-958d-227927f52356","class":null,"textAlign":"left"},"content":[{"type":"text","text":"Cell cultures fail through overgrowth, poor attachment, contamination and environmental and media changes. These failure are currently found after inspection from human intervention days after a culture begins or when media changes are required. [1] This is a large loss to studies as it wastes cultured cells, media, consumable but most of all the scientists time as weekends are often required to inspect cells at both the academic and industrial level. With "}]},{"type":"heading","attrs":{"id":"1504af7b-fd4b-47b2-96fe-7947c6b08e82","data-toc-id":"1504af7b-fd4b-47b2-96fe-7947c6b08e82","textAlign":"left","level":3},"content":[{"type":"text","text":"Hypotheses"}]},{"type":"paragraph","attrs":{"id":"69cbe293-7114-4615-8090-415936addfbe","class":null,"textAlign":"left"},"content":[{"type":"hardBreak"},{"type":"text","text":"H1 . Can ML algorithms determine cell responses before human intervention is required. "}]},{"type":"paragraph","attrs":{"id":"f13d3a01-e378-4b8b-b2b1-1c4a831b8751","class":null,"textAlign":"left"},"content":[{"type":"text","text":"H2. Can a model trained from the first 24 hours predict which well have failed before 96 hours have passed or become visible to assay developmental scientists."}]},{"type":"heading","attrs":{"id":"5db21052-1080-4fc0-bce1-4b1a0690b84d","data-toc-id":"5db21052-1080-4fc0-bce1-4b1a0690b84d","textAlign":"left","level":2},"content":[{"type":"text","text":"3. Methods"}]},{"type":"heading","attrs":{"id":"39c552e4-4ee6-4f4f-a880-2b998198760a","data-toc-id":"39c552e4-4ee6-4f4f-a880-2b998198760a","textAlign":"left","level":3},"content":[{"type":"text","text":"Participants/Sample"}]},{"type":"paragraph","attrs":{"id":"e290afab-d7f3-4378-85a5-d382071a1f9a","class":null,"textAlign":"justify"},"content":[{"type":"hardBreak"},{"type":"text","text":"The target sample will be HEK293 Mammalian cells. Cell growth wells contaminated, failed and successful will provided. Cell lines will be purchased from the ATTC and studied in within 4 different independent experiments seeded on different days. The experiment will consist of 1. Healthy seeded at recommended density given the low doses of nutrients. 2. Healthy cells seeded at the recommended density and the proper amount of nutrients. 3. Healthy Cells given higher doses of nutrients. 4. Healthy cells seeded at the recommended density and proper amount of nutrients spiked with microbial contamination. "},{"type":"hardBreak"},{"type":"text","text":"Temperatures will remain consistent throughout each of the different experimental procedures. lmaging will occur with a bottom up microscope at given times of 1,2,3, 5 and 7 days. Incubator lapse image videos will also be provided to the licensed CUI Fund Pilot. "}]},{"type":"heading","attrs":{"id":"62b8f3d6-c707-4b23-8272-1826e8d107db","data-toc-id":"62b8f3d6-c707-4b23-8272-1826e8d107db","textAlign":"left","level":3},"content":[{"type":"text","text":"Materials and Procedures"}]},{"type":"paragraph","attrs":{"id":"8a47998c-8c8b-41e4-8d46-af85adfa9109","class":null,"textAlign":"left"},"content":[{"type":"hardBreak"},{"type":"text","marks":[{"type":"bold"}],"text":"Cell culture."},{"type":"text","text":" Cells grow in DMEM (high glucose) with 10% fetal bovine serum at 37 °C and 5% CO2 and are passaged at 70–90% confluence with trypsin-EDTA. Antibiotics are omitted from one passage before seeding onward, so that contamination is not masked. Medium and serum lot numbers are recorded."},{"type":"hardBreak"},{"type":"text","marks":[{"type":"bold"}],"text":"Plate setup."},{"type":"text","text":" Healthy wells are seeded at different densities as specified in the Sample section density. Condition positions follow a pre-generated randomized layout. Plates wells use a breathable sealing film and sit in a closed secondary container in the incubator; they are autoclaved after the final image. ( Or left to continue growing until a failure mode occurs to increase the amount of cellular data.)"},{"type":"hardBreak"},{"type":"text","marks":[{"type":"bold"}],"text":"Imaging."},{"type":"text","text":" All images are taken on an inverted phase-contrast microscope. Exposure, gain and light intensity will be fixed. No plate will spend more than 10 minutes out of the incubator per session. The time out of the incubator will be logged. All images will be taken with the microscopy fields and 5 images will be taken per well."}]},{"type":"paragraph","attrs":{"id":"f028706d-fc57-49f0-b5aa-bc84e2308233","class":null,"textAlign":"left"},"content":[{"type":"text","marks":[{"type":"bold"}],"text":"End-point measurements."},{"type":"text","text":" At 96 h, the plate is photographed against a color reference card to record medium color. Every well is then harvested with its floating cells and counted with trypan blue on a hemocytometer, in duplicate."},{"type":"hardBreak"},{"type":"text","marks":[{"type":"bold"}],"text":"Annotations for the dataset. "},{"type":"text","text":"Every image will be listed with the experiment and plate, condition, time, passage. "},{"type":"hardBreak"}]},{"type":"heading","attrs":{"id":"5441ad88-0fa0-4c6d-94a3-63a810a56bbd","data-toc-id":"5441ad88-0fa0-4c6d-94a3-63a810a56bbd","textAlign":"left","level":3},"content":[{"type":"text","text":"Planned Analyses"}]},{"type":"paragraph","attrs":{"id":"a8174e24-da95-4fd2-8514-a83f2773a60d","class":null,"textAlign":"left"},"content":[{"type":"hardBreak"},{"type":"text","text":"In our experimental pipeline analysis the raw images will be provided archived and unchanged. Two Algorithms will be built and compared. One will consist of an anomaly detector, learning the cells growth curve and spread of the features. "}]},{"type":"heading","attrs":{"id":"16fe8d55-579c-442d-b1fe-4daddd0cdabc","data-toc-id":"16fe8d55-579c-442d-b1fe-4daddd0cdabc","textAlign":"left","level":3},"content":[{"type":"text","text":"Ethics and Data Management"}]},{"type":"paragraph","attrs":{"id":"90a7cdea-14b2-4857-8704-e5e32d43dae3","class":null,"textAlign":"left"},"content":[{"type":"text","text":"This study uses an established and commercially available distributed human cell line and involved no patients, patient identifiable data or animals. "},{"type":"hardBreak"},{"type":"text","text":"Data management will consist of seperate raw images, kept under lock and key. Admin privileges are only allowed to view the working drive. "}]},{"type":"heading","attrs":{"id":"e32a1c94-5782-4da9-afe1-315a939fe81e","data-toc-id":"e32a1c94-5782-4da9-afe1-315a939fe81e","textAlign":"left","level":2},"content":[{"type":"text","text":"5. Budget"}]},{"type":"paragraph","attrs":{"id":"c835f7ed-c2d5-4a48-b857-e951a9188aac","class":null,"textAlign":"left"},"content":[{"type":"text","text":"Provide an itemized list of anticipated costs (materials, personnel, etc.). These are not binding and are able to be adjusted based on the results of the pilot data."}]},{"type":"heading","attrs":{"id":"d9c90060-9401-49ae-af24-d2ec4c002f31","data-toc-id":"d9c90060-9401-49ae-af24-d2ec4c002f31","textAlign":"left","level":2},"content":[{"type":"text","text":"6. References"}]},{"type":"paragraph","attrs":{"id":"21c387e9-720d-45f2-92f0-fd75d7d84e0b","class":null,"textAlign":"left"},"content":[{"type":"text","text":"List all references used in the manuscript using Nature style (e.g. numbered in-text citations and alphabetical references at the end of the manuscript)."}]}]}4d:T1d4a,

Early Automated Detection of Cell Culture Failure from Time-Lapse Imaging.

1. Abstract

This study will consists of an incubated mammalian cell type over their weekly growth period. The purpose of the study is to determine whether ML models can determine whether cell growth will fail or succeed in a rapid timeline. The work will be carried out by Andre Childs (PI) at the Dallas Maker Space’s cell culture laboratory. The duration of the project will be last 12 weeks from funding to delivery. Raw images with annotations, protocols and processing logs will be licensed to the CUI Fund Pilot under written agreement and archived openly.

2. Introduction

Cell cultures fail through overgrowth, poor attachment, contamination and environmental and media changes. These failure are currently found after inspection from human intervention days after a culture begins or when media changes are required. [1] This is a large loss to studies as it wastes cultured cells, media, consumable but most of all the scientists time as weekends are often required to inspect cells at both the academic and industrial level. With

Hypotheses

H1 . Can ML algorithms determine cell responses before human intervention is required.

H2. Can a model trained from the first 24 hours predict which well have failed before 96 hours have passed or become visible to assay developmental scientists.

3. Methods

Participants/Sample

The target sample will be HEK293 Mammalian cells. Cell growth wells contaminated, failed and successful will provided. Cell lines will be purchased from the ATTC and studied in within 4 different independent experiments seeded on different days. The experiment will consist of 1. Healthy seeded at recommended density given the low doses of nutrients. 2. Healthy cells seeded at the recommended density and the proper amount of nutrients. 3. Healthy Cells given higher doses of nutrients. 4. Healthy cells seeded at the recommended density and proper amount of nutrients spiked with microbial contamination.
Temperatures will remain consistent throughout each of the different experimental procedures. lmaging will occur with a bottom up microscope at given times of 1,2,3, 5 and 7 days. Incubator lapse image videos will also be provided to the licensed CUI Fund Pilot.

Materials and Procedures

Cell culture. Cells grow in DMEM (high glucose) with 10% fetal bovine serum at 37 °C and 5% CO2 and are passaged at 70–90% confluence with trypsin-EDTA. Antibiotics are omitted from one passage before seeding onward, so that contamination is not masked. Medium and serum lot numbers are recorded.
Plate setup. Healthy wells are seeded at different densities as specified in the Sample section density. Condition positions follow a pre-generated randomized layout. Plates wells use a breathable sealing film and sit in a closed secondary container in the incubator; they are autoclaved after the final image. ( Or left to continue growing until a failure mode occurs to increase the amount of cellular data.)
Imaging. All images are taken on an inverted phase-contrast microscope. Exposure, gain and light intensity will be fixed. No plate will spend more than 10 minutes out of the incubator per session. The time out of the incubator will be logged. All images will be taken with the microscopy fields and 5 images will be taken per well.

End-point measurements. At 96 h, the plate is photographed against a color reference card to record medium color. Every well is then harvested with its floating cells and counted with trypan blue on a hemocytometer, in duplicate.
Annotations for the dataset. Every image will be listed with the experiment and plate, condition, time, passage.

Planned Analyses

In our experimental pipeline analysis the raw images will be provided archived and unchanged. Two Algorithms will be built and compared. One will consist of an anomaly detector, learning the cells growth curve and spread of the features.

Ethics and Data Management

This study uses an established and commercially available distributed human cell line and involved no patients, patient identifiable data or animals.
Data management will consist of seperate raw images, kept under lock and key. Admin privileges are only allowed to view the working drive.

5. Budget

Provide an itemized list of anticipated costs (materials, personnel, etc.). These are not binding and are able to be adjusted based on the results of the pilot data.

6. References

List all references used in the manuscript using Nature style (e.g. numbered in-text citations and alphabetical references at the end of the manuscript).

4e:T1d4a,

Early Automated Detection of Cell Culture Failure from Time-Lapse Imaging.

1. Abstract

2. Introduction

Hypotheses

H1 . Can ML algorithms determine cell responses before human intervention is required.

H2. Can a model trained from the first 24 hours predict which well have failed before 96 hours have passed or become visible to assay developmental scientists.

3. Methods

Participants/Sample

Materials and Procedures

Planned Analyses

Ethics and Data Management

5. Budget

Provide an itemized list of anticipated costs (materials, personnel, etc.). These are not binding and are able to be adjusted based on the results of the pilot data.

6. References

List all references used in the manuscript using Nature style (e.g. numbered in-text citations and alphabetical references at the end of the manuscript).

4f:T1220,Early Automated Detection of Cell Culture Failure from Time-Lapse Imaging. 1. Abstract This study will consists of an incubated mammalian cell type over their weekly growth period. The purpose of the study is to determine whether ML models can determine whether cell growth will fail or succeed in a rapid timeline. The work will be carried out by Andre Childs (PI) at the Dallas Maker Space’s cell culture laboratory. The duration of the project will be last 12 weeks from funding to delivery. Raw images with annotations, protocols and processing logs will be licensed to the CUI Fund Pilot under written agreement and archived openly. 2. Introduction Cell cultures fail through overgrowth, poor attachment, contamination and environmental and media changes. These failure are currently found after inspection from human intervention days after a culture begins or when media changes are required. [1] This is a large loss to studies as it wastes cultured cells, media, consumable but most of all the scientists time as weekends are often required to inspect cells at both the academic and industrial level. With Hypotheses H1 . Can ML algorithms determine cell responses before human intervention is required. H2. Can a model trained from the first 24 hours predict which well have failed before 96 hours have passed or become visible to assay developmental scientists. 3. Methods Participants/Sample The target sample will be HEK293 Mammalian cells. Cell growth wells contaminated, failed and successful will provided. Cell lines will be purchased from the ATTC and studied in within 4 different independent experiments seeded on different days. The experiment will consist of 1. Healthy seeded at recommended density given the low doses of nutrients. 2. Healthy cells seeded at the recommended density and the proper amount of nutrients. 3. Healthy Cells given higher doses of nutrients. 4. Healthy cells seeded at the recommended density and proper amount of nutrients spiked with microbial contamination. Temperatures will remain consistent throughout each of the different experimental procedures. lmaging will occur with a bottom up microscope at given times of 1,2,3, 5 and 7 days. Incubator lapse image videos will also be provided to the licensed CUI Fund Pilot. Materials and Procedures Cell culture. Cells grow in DMEM (high glucose) with 10% fetal bovine serum at 37 °C and 5% CO2 and are passaged at 70–90% confluence with trypsin-EDTA. Antibiotics are omitted from one passage before seeding onward, so that contamination is not masked. Medium and serum lot numbers are recorded. Plate setup. Healthy wells are seeded at different densities as specified in the Sample section density. Condition positions follow a pre-generated randomized layout. Plates wells use a breathable sealing film and sit in a closed secondary container in the incubator; they are autoclaved after the final image. ( Or left to continue growing until a failure mode occurs to increase the amount of cellular data.) Imaging. All images are taken on an inverted phase-contrast microscope. Exposure, gain and light intensity will be fixed. No plate will spend more than 10 minutes out of the incubator per session. The time out of the incubator will be logged. All images will be taken with the microscopy fields and 5 images will be taken per well. End-point measurements. At 96 h, the plate is photographed against a color reference card to record medium color. Every well is then harvested with its floating cells and counted with trypan blue on a hemocytometer, in duplicate. Annotations for the dataset. Every image will be listed with the experiment and plate, condition, time, passage. Planned Analyses In our experimental pipeline analysis the raw images will be provided archived and unchanged. Two Algorithms will be built and compared. One will consist of an anomaly detector, learning the cells growth curve and spread of the features. Ethics and Data Management This study uses an established and commercially available distributed human cell line and involved no patients, patient identifiable data or animals. Data management will consist of seperate raw images, kept under lock and key. Admin privileges are only allowed to view the working drive. 5. Budget Provide an itemized list of anticipated costs (materials, personnel, etc.). These are not binding and are able to be adjusted based on the results of the pilot data. 6. References List all references used in the manuscript using Nature style (e.g. numbered in-text citations and alphabetical references at the end of the manuscript).1d:["$","$L46",null,{"children":["$","$L4",null,{"topBanner":["$","$L47",null,{"work":{"id":32994,"type":"$undefined","contentType":"preregistration","title":"Early Automated Detection of Cell Culture Failure from Time-Lapse Imaging.","slug":"early-automated-detection-of-cell-culture-failure-from-time-lapse-imaging","createdDate":"2026-10-01T05:01:50.793856Z","createdByUserId":204287,"updatedDate":"2026-10-07T16:47:56.026048Z","publishedDate":"2026-10-01T05:01:50.793856Z","authors":[{"authorProfile":{"id":8690589,"fullName":"André Childs","firstName":"André","lastName":"Childs","profileImage":"","headline":"","profileUrl":"/author/8690589","user":{"id":0,"email":"","firstName":"","lastName":"","fullName":"Unknown User","isVerified":false,"authorProfile":"$undefined","balance":0,"lockedBalance":0,"promotionalBalance":0,"fundingCredits":0,"balances":"$undefined","totalRsc":0,"hasCompletedOnboarding":false,"createdDate":"$undefined","moderator":false,"editorOfHubs":"$undefined","isModerator":false,"isFunder":false,"referralCode":"$undefined","authProvider":"$undefined","isStakingOptedIn":false,"balanceHistory":0,"raw":204287},"description":"$undefined","createdDate":"2026-07-17T22:31:09.654901Z","education":[{"id":8213,"name":"University of Central Florida","year":{"label":"2024","value":"2024"},"major":"Material Science and Engineering","degree":{"label":"Doctor of Philosophy","value":"PhD"},"summary":"Material Science and Engineering PhD '24, University of Central Florida","is_public":true,"university":{"id":8213,"city":"Orlando","name":"University of Central Florida","state":"FL","country":"United States","updated_date":"2020-11-24T21:44:27.626414Z"}}],"twitter":"$undefined","facebook":"$undefined","linkedin":"https://www.linkedin.com/in/andrechilds/","googleScholar":"$undefined","orcidId":"https://orcid.org/0009-0004-6002-9945","isOrcidConnected":true,"isClaimed":true,"userId":"$undefined","isVerified":true,"isHubEditor":false,"editorOfHubs":[],"raw":{"id":8690589,"is_public":true,"is_removed":false,"is_removed_date":null,"user":204287,"first_name":"André","last_name":"Childs","created_date":"2026-07-17T22:31:09.654901Z","updated_date":"2026-07-18T20:15:25.944314Z","description":"","h_index":3,"i10_index":3,"profile_image":null,"author_score":0,"university":null,"orcid_id":"https://orcid.org/0009-0004-6002-9945","openalex_ids":["https://openalex.org/A5042270030"],"education":"$1d:props:children:props:topBanner:props:work:authors:0:authorProfile:education","headline":"","facebook":null,"twitter":"","linkedin":"https://www.linkedin.com/in/andrechilds/","google_scholar":"","claimed":true,"country_code":null,"created_source":"RESEARCHHUB","last_full_fetch_from_openalex":null,"two_year_mean_citedness":171,"merged_with_author":null,"added_as_editor_date":null,"is_hub_editor_of":[],"is_hub_editor":false,"num_posts":1,"is_orcid_connected":true,"reputation":100,"suspended_status":{"probable_spammer":false,"is_suspended":false},"total_score":null,"wallet":null,"is_verified":true}},"isCorresponding":false,"position":"middle","raw":"$1d:props:children:props:topBanner:props:work:authors:0:authorProfile:raw"}],"abstract":"$48","doi":null,"journal":"$undefined","formats":[],"license":"$undefined","pdfCopyrightAllowsDisplay":true,"figures":[],"versions":[],"metrics":{"votes":4,"adjustedScore":"$undefined","comments":3,"saves":0,"reviewScore":3,"reviews":1,"earned":0,"views":0},"unifiedDocumentId":9367319,"registeredReportId":null,"postType":"PREREGISTRATION","fundraise":"$undefined","note":{"id":58296,"access":"WORKSPACE","organization":{"id":199464,"memberCount":1,"userPermission":"$undefined","createdDate":"2026-07-17T22:31:09.855447Z","coverImage":null,"description":"","name":"André Childs's Org","slug":"andre-childss-org"},"createdDate":"2026-10-01T03:51:27.390883Z","updatedDate":"2026-10-01T05:01:49.155635Z","title":"Early Automated Detection of Cell Culture Failure from Time-Lapse Imaging.","isRemoved":false,"post":{"id":32994,"slug":"early-automated-detection-of-cell-culture-failure-from-time-lapse-imaging","contentType":"preregistration","documentType":"PREREGISTRATION","moderationStatus":"APPROVED","fundraise":{"id":1208,"amountRaised":{"usd":0,"rsc":0},"goalAmount":{"usd":5000,"rsc":93643.52017724006},"contributors":{"numContributors":0,"topContributors":[]},"createdBy":{"id":0,"email":"","firstName":"","lastName":"","fullName":"Unknown User","isVerified":false,"authorProfile":"$undefined","balance":0,"lockedBalance":0,"totalRsc":0,"hasCompletedOnboarding":false,"moderator":false,"raw":null,"isModerator":false,"isFunder":false,"authProvider":"$undefined","isStakingOptedIn":false,"balanceHistory":0},"status":"OPEN","goalCurrency":"USD","startDate":"2026-10-01T05:01:50.845556Z","endDate":"2026-11-30T06:01:50.679000Z","createdDate":"","updatedDate":"","postId":"$undefined","postTitle":"$undefined","postSlug":"$undefined","postImage":null,"reviewMetrics":"$undefined","nonprofit":"$undefined","raw":{"id":1208,"amount_raised":{"usd":0,"rsc":0},"goal_amount":{"usd":5000,"rsc":93643.52017724006},"contributors":{"total":0,"top":[]},"status":"OPEN","goal_currency":"USD","start_date":"2026-10-01T05:01:50.845556Z","end_date":"2026-11-30T06:01:50.679000Z"}},"grant":"$undefined","doi":null,"authors":[{"authorId":8690589,"userId":204287,"name":"André Childs","raw":{"id":8690589,"first_name":"André","last_name":"Childs","user":204287}}],"contacts":"$undefined","image":"https://storage.prod.researchhub.com/uploads/posts/users/204287/86afa818-bedd-458f-bded-a2632210c7f8/screenshot-2026-09-30-at-11.png","raw":{"id":32994,"status":"APPROVED","authors":["$1d:props:children:props:topBanner:props:work:note:post:authors:0:raw"],"unified_document":{"fundraise":"$1d:props:children:props:topBanner:props:work:note:post:fundraise:raw","grant":null},"image_url":"https://storage.prod.researchhub.com/uploads/posts/users/204287/86afa818-bedd-458f-bded-a2632210c7f8/screenshot-2026-09-30-at-11.png","document_type":"PREREGISTRATION","slug":"early-automated-detection-of-cell-culture-failure-from-time-lapse-imaging","doi":null}},"documentType":"PREREGISTRATION","proposalId":null,"registeredReportPrefill":null,"image":"uploads/posts/users/204287/86afa818-bedd-458f-bded-a2632210c7f8/screenshot-2026-09-30-at-11.png","previewImage":"https://storage.prod.researchhub.com/uploads/posts/users/204287/86afa818-bedd-458f-bded-a2632210c7f8/screenshot-2026-09-30-at-11.png","authors":[{"authorId":8690589,"userId":204287,"name":"André Childs","raw":{"id":8690589,"first_name":"André","last_name":"Childs","profile_image":null,"user":204287}}],"grantSettings":null,"preregistrationSettings":{"goalAmount":"5000.00","durationDays":null,"isPublic":true,"nonprofit":null,"raw":{"duration_days":null,"goal_amount":"5000.00","goal_currency":"USD","is_public":true,"nonprofit_details":null,"nonprofit_id":null}},"selectedGrant":{"id":"343","shortTitle":"Give New Meaning to Existing Live-Cell Imaging Data","imageUrl":"https://storage.prod.researchhub.com/uploads/posts/users/202314/bb92f454-f415-4183-86f4-1f38b2c6b163/screenshot-2026-07-16-at-17.png","fundingAmount":5000,"organization":"Ben Noach","applicationVisibility":"PUBLIC","raw":{"id":343,"amount":{"usd":5000,"rsc":93643.52017724006,"formatted":"5,000.00 USD"},"image_url":"https://storage.prod.researchhub.com/uploads/posts/users/202314/bb92f454-f415-4183-86f4-1f38b2c6b163/screenshot-2026-07-16-at-17.png","title":"Give New Meaning to Existing Live-Cell Imaging Data","organization":"Ben Noach","short_title":null,"application_visibility":"PUBLIC"}},"raw":{"id":58296,"access":"WORKSPACE","authors":["$1d:props:children:props:topBanner:props:work:note:authors:0:raw"],"grant_settings":null,"latest_version":{"id":588260,"src":null,"created_date":"2026-10-01T05:01:37.985010Z","created_via":"editor","plain_text":"$49","json":"$4a","created_by":204287,"note":58296,"parent_version":null},"organization":{"id":199464,"member_count":1,"user_permission":null,"created_date":"2026-07-17T22:31:09.855447Z","updated_date":"2026-07-17T22:31:09.855458Z","cover_image":null,"description":"","name":"André Childs's Org","note_created":false,"org_type":"PERSONAL","slug":"andre-childss-org","user":204287},"post":"$1d:props:children:props:topBanner:props:work:note:post:raw","preregistration_settings":"$1d:props:children:props:topBanner:props:work:note:preregistrationSettings:raw","registered_report_prefill":null,"selected_grant":343,"selected_grant_details":"$1d:props:children:props:topBanner:props:work:note:selectedGrant:raw","unified_document":{"id":9367317,"is_removed":false},"created_date":"2026-10-01T03:51:27.390883Z","updated_date":"2026-10-01T05:01:49.155635Z","document_type":"PREREGISTRATION","image":"uploads/posts/users/204287/86afa818-bedd-458f-bded-a2632210c7f8/screenshot-2026-09-30-at-11.png","preview_img":"https://storage.prod.researchhub.com/uploads/posts/users/204287/86afa818-bedd-458f-bded-a2632210c7f8/screenshot-2026-09-30-at-11.png","title":"Early Automated Detection of Cell Culture Failure from Time-Lapse Imaging.","created_by":204287},"content":null,"versionId":588260,"versionDate":"2026-10-01T05:01:37.985010Z","plainText":"$4b","contentJson":"$4c","versionCreatedVia":"editor"},"previewContent":"$4d","contentUrl":"https://storage.prod.researchhub.com/uploads/post_discussion/2026/10/01/RH-POST-PREREGISTRATION-USER-204287.txt","tips":[],"peerReviews":[{"id":19307,"createdBy":{"id":197732,"authorProfile":{"id":8623156,"fullName":"AI Review","profileImage":"https://storage.prod.researchhub.com/uploads/author_profile_images/2026/07/02/blob_Yfubd6a","isVerified":false}},"score":3,"createdDate":"2026-10-01T05:02:56.674330Z","contentPreview":"$undefined","isAssessed":true}],"image":"https://storage.prod.researchhub.com/uploads/posts/users/204287/86afa818-bedd-458f-bded-a2632210c7f8/screenshot-2026-09-30-at-11.png","moderationStatus":"APPROVED","raw":{"score":4,"user_flag":null,"authors":["$1d:props:children:props:topBanner:props:work:authors:0:authorProfile:raw"],"id":32994,"created_by":{"id":204287,"author_profile":{"id":8690589,"is_public":true,"is_removed":false,"is_removed_date":null,"user":204287,"first_name":"André","last_name":"Childs","created_date":"2026-07-17T22:31:09.654901Z","updated_date":"2026-07-18T20:15:25.944314Z","description":"","h_index":3,"i10_index":3,"profile_image":null,"author_score":0,"university":null,"orcid_id":"https://orcid.org/0009-0004-6002-9945","openalex_ids":["https://openalex.org/A5042270030"],"education":[{"id":8213,"name":"University of Central Florida","year":{"label":"2024","value":"2024"},"major":"Material Science and Engineering","degree":{"label":"Doctor of Philosophy","value":"PhD"},"summary":"Material Science and Engineering PhD '24, University of Central Florida","is_public":true,"university":{"id":8213,"city":"Orlando","name":"University of Central Florida","state":"FL","country":"United States","updated_date":"2020-11-24T21:44:27.626414Z"}}],"headline":"","facebook":null,"twitter":"","linkedin":"https://www.linkedin.com/in/andrechilds/","google_scholar":"","claimed":true,"country_code":null,"created_source":"RESEARCHHUB","last_full_fetch_from_openalex":null,"two_year_mean_citedness":171,"merged_with_author":null,"added_as_editor_date":null,"is_hub_editor_of":[],"is_hub_editor":false,"num_posts":1,"is_orcid_connected":true,"reputation":100,"suspended_status":{"probable_spammer":false,"is_suspended":false},"total_score":null,"wallet":null,"is_verified":true},"balance":null,"balances":null,"created_date":"2026-07-17T22:31:09.610094Z","has_seen_first_coin_modal":false,"has_seen_orcid_connect_modal":false,"is_funder":false,"is_staking_opted_in":true,"is_suspended":false,"probable_spammer":false,"is_verified":true,"moderator":false,"reputation":100,"updated_date":"2026-07-17T22:31:09.610127Z","upload_tutorial_complete":false,"hub_rep":null,"time_rep":null},"created_date":"2026-10-01T05:01:50.793856Z","discussion_count":3,"document_type":"PREREGISTRATION","doi":null,"editor_type":"CK_EDITOR","full_markdown":"$4e","grants":[{"id":343,"short_title":null,"status":"OPEN","organization":"Ben Noach","amount":"5000.00","currency":"USD","post_id":32623,"image_url":"https://storage.prod.researchhub.com/uploads/posts/users/202314/bb92f454-f415-4183-86f4-1f38b2c6b163/screenshot-2026-07-16-at-17.png","title":"Give New Meaning to Existing Live-Cell Imaging Data","applicant_count":1,"application_visibility":"PUBLIC","application_id":342,"created_by":{"id":202314},"funding_pool":{"id":265,"amount_holding":{"rsc":0,"usd":0},"amount_distributed":{"rsc":0,"usd":0},"amount_raised":{"rsc":0,"usd":0},"status":"OPEN"},"proposal":{"unified_document_id":9367319,"ai_peer_review":{"id":231,"unified_document_id":9367319,"grant_id":343,"created_by_id":null,"status":"completed","overall_rating":"adequate","overall_rationale":"Relevant to the RFP's dataset focus and addresses a real pain point, but the proposal is underdeveloped: missing budget, missing references, thin ML plan, and mismatch with the RFP's preference for existing annotated datasets.","overall_score_numeric":3,"result_data":{"categories":{"overall_impact":{"items":{"rigor":{"decision":"No","justification":"Analysis plan is a single incomplete sentence mentioning an anomaly detector and a second unspecified algorithm. No metrics, train/test split, cross-validation, or baselines are described."},"novelty":{"decision":"Partial","justification":"Early ML-based detection of culture failure is a reasonable but not novel direction; prior work exists on time-lapse ML for microbial and iPSC cultures. The specific framing around HEK293 failure modes with contamination spike-ins is a modest contribution."},"field_impact":{"decision":"Partial","justification":"If the dataset is well-annotated and openly licensed, it could be useful to the CUI pilot and similar efforts, but the small scale and single cell line constrain generalizability."},"reproducibility":{"decision":"Partial","justification":"Culture and imaging protocols are reasonably specified (media, CO2, exposure fixed, logged time out of incubator), and raw images will be archived. However, references are placeholder text and budget is not itemized, limiting reproducibility of the full study."}},"score":2,"rationale":"Addresses a practical and widely felt problem in cell culture, but the pilot's scope (one cell line, 4 experiments, 12 weeks) and underspecified analysis limit likely field impact. Reproducibility is partially supported by protocol details but undermined by missing budget, references, and model specification."},"rigor_and_feasibility":{"items":{"methodology":{"decision":"Partial","justification":"Culture, antibiotic washout, phase-contrast with fixed exposure, trypan blue counts in duplicate, and color-reference endpoint photos are appropriate. However, the ML methodology is one incomplete sentence with no architectures, features, metrics, or validation plan."},"study_design":{"decision":"Partial","justification":"Four conditions (low nutrients, normal, high nutrients, contamination spike) across 4 independent seedings is a sensible pilot design, but replication per condition, well counts, and statistical power are not stated. Imaging at days 1,2,3,5,7 may not support a 24h prediction endpoint cleanly."},"team_qualifications":{"decision":"Partial","justification":"PI's public record (OpenAlex h-index 3, biosensor/electrochemical sensing background, UCF affiliation) shows relevant biomedical/engineering experience but no listed prior cell-culture ML publications. No co-investigators or ML collaborators named."},"research_environment":{"decision":"Partial","justification":"Dallas Makerspace cell culture lab is cited; makerspaces can support basic culture work, but incubator stability, sterile technique infrastructure, and long-duration imaging capability are not documented in the proposal."},"timeline_feasibility":{"decision":"Partial","justification":"12 weeks to run 4 independent experiments with weekly growth cycles, perform annotations, and build/compare two ML models is tight. Feasible for data collection but likely rushed for annotation and modeling."},"budget_appropriateness_justification":{"decision":"No","justification":"Budget section is literally the template instruction ('Provide an itemized list of anticipated costs') with no line items, totals, or justification. Cannot be assessed against the EUR 5,000 shared pilot fund."}},"score":3,"rationale":"Wet-lab design is reasonable with randomized layouts, antibiotic washout, and standard end-point counts, but the imaging cadence (days 1, 2, 3, 5, 7) is coarse for a 24h prediction target, the ML analysis is severely underspecified, the budget is a template placeholder, and the environment is a community makerspace rather than a documented research lab."},"additional_review_criteria":{"items":{"ai_use_disclosed":{"decision":"Partial","justification":"ML models are the core method and are described as such, but there is no disclosure about use of AI tools in writing the proposal or specifics of model provenance."},"open_science_adherence":{"decision":"Yes","justification":"Raw images, annotations, protocols, and processing logs will be licensed to the CUI Fund Pilot and 'archived openly,' aligning with the RFP's data-sharing orientation."},"human_or_animal_protections":{"decision":"Yes","justification":"Study uses an established commercial human cell line (HEK293) with no patient-identifiable data or animals, as stated in the Ethics section."},"conflicts_of_interest_disclosed":{"decision":"No","justification":"No conflict-of-interest statement is included in the proposal text."},"resubmission_critiques_addressed":{"decision":"Partial","justification":"No indication this is a resubmission; criterion is not applicable but marked Partial per no-N/A policy with neutral justification."}},"score":3,"rationale":"No human/animal concerns (established HEK293 line). Open sharing under written license is proposed. AI use and conflicts are not explicitly discussed, and resubmission does not apply."},"importance_significance_innovation":{"items":{"work_novelty":{"decision":"Partial","justification":"Combining contamination spike-ins with nutrient perturbations in a labeled HEK293 time-lapse set is a useful dataset contribution, though the ML methodology is not differentiated from existing approaches."},"advances_knowledge":{"decision":"Partial","justification":"A small, well-annotated dataset could advance method development, but the pilot alone is unlikely to produce generalizable models given sample size and single cell line."},"hypothesis_strength":{"decision":"Partial","justification":"H1 and H2 are posed as yes/no questions rather than falsifiable hypotheses with effect sizes or decision thresholds. H2's claim that 24h data can predict 96h outcomes is testable but not quantified."},"question_importance":{"decision":"Yes","justification":"Culture failure causes substantial waste of reagents, cells, and researcher time; early detection has clear practical value in academic and industrial labs."}},"score":4,"rationale":"The underlying question - early automated detection of culture failure - is important and clearly motivated by real costs to labs. The hypotheses are stated but imprecise, and novelty is incremental relative to published time-lapse ML work."}},"overall_rating":"adequate","overall_summary":"The proposal describes a 12-week pilot to generate a HEK293 time-lapse phase-contrast imaging dataset across healthy, nutrient-perturbed, and contaminated conditions, with the goal of training ML models (anomaly detection and a predictive classifier) to flag culture failure within 24 hours. The idea is well-aligned with the CUI Fund call for annotated 2D mammalian live-cell microscopy datasets, and the experimental conditions map cleanly to failure modes of interest. However, the proposal has important gaps: the RFP explicitly seeks existing annotated datasets, while this project proposes to generate new data; the budget section is a template placeholder with no itemization; references are not provided; the ML analysis plan is only a sentence and lacks baselines, metrics, and validation strategy; and the imaging cadence (days 1-7) may be too coarse to support a 24-hour prediction claim. Team expertise in biosensors is relevant but prior cell-culture ML work is not evidenced.","overall_rationale":"Relevant to the RFP's dataset focus and addresses a real pain point, but the proposal is underdeveloped: missing budget, missing references, thin ML plan, and mismatch with the RFP's preference for existing annotated datasets.","overall_score_numeric":3},"error_message":"","progress":100,"current_step":"Complete","llm_model":"us.anthropic.claude-opus-4-7","processing_time":42.643226733998745,"created_date":"2026-10-01T05:02:13.809447Z","updated_date":"2026-10-01T05:02:56.517762Z","editorial_feedback":null,"key_insight":{"id":153,"status":"completed","tldr":"A 12-week pilot to generate HEK293 time-lapse phase-contrast images across healthy, nutrient-perturbed, and contaminated wells, then train ML models to flag culture failure early. Fit to the RFP is partial: the call seeks existing annotated datasets, while this proposes new data collection. Key gaps include a placeholder budget, missing references, and a one-sentence ML plan without metrics or baselines.","error_message":"","llm_model":"us.anthropic.claude-opus-4-7","processing_time":9.680480675000581,"created_date":"2026-10-01T05:02:56.751435Z","updated_date":"2026-10-01T05:03:06.435562Z","items":[{"id":3165,"item_type":"strength","label":"Relevant Problem","description":"Early detection of culture failure addresses real reagent, cell, and labor waste in academic and industrial labs, with clear practical value.","order":0,"created_date":"2026-10-01T05:03:06.438570Z","updated_date":"2026-10-01T05:03:06.438578Z"},{"id":3166,"item_type":"strength","label":"Open Data Plan","description":"Raw images, annotations, protocols, and logs will be licensed to the CUI pilot and archived openly, aligning with the RFP's data-sharing goals.","order":1,"created_date":"2026-10-01T05:03:06.438597Z","updated_date":"2026-10-01T05:03:06.438600Z"},{"id":3167,"item_type":"strength","label":"Sensible Conditions","description":"Four conditions spanning nutrient perturbation and contamination spike-ins map cleanly to common failure modes and yield interpretable labels.","order":2,"created_date":"2026-10-01T05:03:06.438610Z","updated_date":"2026-10-01T05:03:06.438612Z"},{"id":3168,"item_type":"strength","label":"Wet-Lab Rigor","description":"Fixed imaging exposure, antibiotic washout, randomized layouts, logged incubator time, and duplicate trypan blue counts support reproducible data collection.","order":3,"created_date":"2026-10-01T05:03:06.438621Z","updated_date":"2026-10-01T05:03:06.438623Z"},{"id":3169,"item_type":"weakness","label":"RFP Misalignment","description":"The call explicitly seeks existing annotated datasets, but this proposal generates new data within the 12-week period, which may fall outside scope.","order":0,"created_date":"2026-10-01T05:03:06.438631Z","updated_date":"2026-10-01T05:03:06.438634Z"},{"id":3170,"item_type":"weakness","label":"Underspecified ML","description":"The analysis is a single incomplete sentence with no architectures, metrics, baselines, train/test split, or validation plan for the two models.","order":1,"created_date":"2026-10-01T05:03:06.438642Z","updated_date":"2026-10-01T05:03:06.438645Z"},{"id":3171,"item_type":"weakness","label":"Missing Budget","description":"Budget section is template placeholder text with no line items or totals, so fit to the shared 5,000 euro pilot fund cannot be assessed.","order":2,"created_date":"2026-10-01T05:03:06.438652Z","updated_date":"2026-10-01T05:03:06.438655Z"},{"id":3172,"item_type":"weakness","label":"Imaging Cadence","description":"Imaging on days 1, 2, 3, 5, 7 is coarse for an H2 claim that 24-hour data predicts 96-hour outcomes, weakening the central prediction test.","order":3,"created_date":"2026-10-01T05:03:06.438663Z","updated_date":"2026-10-01T05:03:06.438666Z"},{"id":3173,"item_type":"weakness","label":"Scale And Setting","description":"One cell line, four seedings, and a community makerspace with no documented ML collaborator or power analysis limit generalizability and feasibility.","order":4,"created_date":"2026-10-01T05:03:06.438674Z","updated_date":"2026-10-01T05:03:06.438676Z"}]}},"application_id":342}}],"image_url":"https://storage.prod.researchhub.com/uploads/posts/users/204287/86afa818-bedd-458f-bded-a2632210c7f8/screenshot-2026-09-30-at-11.png","is_latest_version":null,"is_removed":false,"is_root_version":true,"note":"$1d:props:children:props:topBanner:props:work:note:raw","peer_reviews":[{"id":19307,"created_by":{"id":197732,"author_profile":{"id":8623156,"first_name":"AI","last_name":"Review","profile_image":"https://storage.prod.researchhub.com/uploads/author_profile_images/2026/07/02/blob_Yfubd6a"},"first_name":"AI","last_name":"Review"},"created_date":"2026-10-01T05:02:56.674330Z","updated_date":"2026-10-01T05:02:56.674342Z","score":3,"is_assessed":true}],"post_src":"https://storage.prod.researchhub.com/uploads/post_discussion/2026/10/01/RH-POST-PREREGISTRATION-USER-204287.txt","preview_img":null,"registered_report_id":null,"renderable_text":"$4f","reviewed_by":null,"reviewed_date":null,"slug":"early-automated-detection-of-cell-culture-failure-from-time-lapse-imaging","status":"APPROVED","title":"Early Automated Detection of Cell Culture Failure from Time-Lapse Imaging. ","unified_document_id":9367319,"unified_document":{"id":9367319,"documents":[{"id":32994,"title":"Early Automated Detection of Cell Culture Failure from Time-Lapse Imaging. ","slug":"early-automated-detection-of-cell-culture-failure-from-time-lapse-imaging"}],"created_by":{"id":204287,"author_profile":{"id":8690589,"first_name":"André","last_name":"Childs","profile_image":null}},"reviews":{"avg":3,"count":1},"is_public":true,"is_removed":false,"document_type":"PREREGISTRATION"},"version_number":1,"updated_date":"2026-10-07T16:47:56.026048Z","bounty_type":null},"aiPeerReview":{"status":"completed","overallRating":null,"keyInsight":{"tldr":"A 12-week pilot to generate HEK293 time-lapse phase-contrast images across healthy, nutrient-perturbed, and contaminated wells, then train ML models to flag culture failure early. Fit to the RFP is partial: the call seeks existing annotated datasets, while this proposes new data collection. Key gaps include a placeholder budget, missing references, and a one-sentence ML plan without metrics or baselines.","items":[{"id":3165,"order":0,"itemType":"strength","label":"Relevant Problem","description":"Early detection of culture failure addresses real reagent, cell, and labor waste in academic and industrial labs, with clear practical value."},{"id":3169,"order":0,"itemType":"weakness","label":"RFP Misalignment","description":"The call explicitly seeks existing annotated datasets, but this proposal generates new data within the 12-week period, which may fall outside scope."},{"id":3166,"order":1,"itemType":"strength","label":"Open Data Plan","description":"Raw images, annotations, protocols, and logs will be licensed to the CUI pilot and archived openly, aligning with the RFP's data-sharing goals."},{"id":3170,"order":1,"itemType":"weakness","label":"Underspecified ML","description":"The analysis is a single incomplete sentence with no architectures, metrics, baselines, train/test split, or validation plan for the two models."},{"id":3167,"order":2,"itemType":"strength","label":"Sensible Conditions","description":"Four conditions spanning nutrient perturbation and contamination spike-ins map cleanly to common failure modes and yield interpretable labels."},{"id":3171,"order":2,"itemType":"weakness","label":"Missing Budget","description":"Budget section is template placeholder text with no line items or totals, so fit to the shared 5,000 euro pilot fund cannot be assessed."},{"id":3168,"order":3,"itemType":"strength","label":"Wet-Lab Rigor","description":"Fixed imaging exposure, antibiotic washout, randomized layouts, logged incubator time, and duplicate trypan blue counts support reproducible data collection."},{"id":3172,"order":3,"itemType":"weakness","label":"Imaging Cadence","description":"Imaging on days 1, 2, 3, 5, 7 is coarse for an H2 claim that 24-hour data predicts 96-hour outcomes, weakening the central prediction test."},{"id":3173,"order":4,"itemType":"weakness","label":"Scale And Setting","description":"One cell line, four seedings, and a community makerspace with no documented ML collaborator or power analysis limit generalizability and feasibility."}]},"raw":"$1d:props:children:props:topBanner:props:work:raw:grants:0:proposal:ai_peer_review"},"linkedGrant":{"id":343,"postId":32623,"title":"Give New Meaning to Existing Live-Cell Imaging Data","shortTitle":null,"organization":"Ben Noach","fundingAmount":5000,"currency":"USD","status":"OPEN","applicationVisibility":"PUBLIC","imageUrl":"https://storage.prod.researchhub.com/uploads/posts/users/202314/bb92f454-f415-4183-86f4-1f38b2c6b163/screenshot-2026-07-16-at-17.png","applicantCount":1,"fundingPool":{"id":265,"status":"OPEN","amountHolding":{"usd":0,"rsc":0},"amountDistributed":{"usd":0,"rsc":0},"amountRaised":{"usd":0,"rsc":0},"contributors":{"total":0,"top":[]}},"createdByUserId":202314,"applicationId":342},"isPublic":true},"metadata":{"id":9367319,"score":4,"metrics":{"votes":4,"comments":3,"saves":0,"reviewScore":3,"conversationComments":3,"reviewComments":1,"bountyComments":0},"fundraising":{"id":1208,"amountRaised":{"usd":0,"rsc":0},"goalAmount":{"usd":5000,"rsc":93643.52017724006},"contributors":{"numContributors":0,"topContributors":[]},"createdBy":{"id":0,"email":"","firstName":"","lastName":"","fullName":"Unknown User","isVerified":false,"authorProfile":"$undefined","balance":0,"lockedBalance":0,"totalRsc":0,"hasCompletedOnboarding":false,"moderator":false,"raw":null,"isModerator":false,"isFunder":false,"authProvider":"$undefined","isStakingOptedIn":false,"balanceHistory":0},"status":"OPEN","goalCurrency":"USD","startDate":"2026-10-01T05:01:50.845556Z","endDate":"2026-11-30T06:01:50.679000Z","createdDate":"","updatedDate":"","postId":"$undefined","postTitle":"$undefined","postSlug":"$undefined","postImage":null,"reviewMetrics":"$undefined","nonprofit":"$undefined","raw":{"id":1208,"amount_raised":{"usd":0,"rsc":0},"goal_amount":{"usd":5000,"rsc":93643.52017724006},"contributors":{"total":0,"top":[]},"status":"OPEN","goal_currency":"USD","start_date":"2026-10-01T05:01:50.845556Z","end_date":"2026-11-30T06:01:50.679000Z"}},"bounties":[],"activeBounties":0,"closedBounties":0},"updatesCount":0,"preTitle":"$L50"}],"rightSidebar":"$L51","children":"$L52"}]}] 54:T147d,

Hypotheses

H1 . Can ML algorithms determine cell responses before human intervention is required.

H2. Can a model trained from the first 24 hours predict which well have failed before 96 hours have passed or become visible to assay developmental scientists.

3. Methods

Participants/Sample

Materials and Procedures

Planned Analyses

Ethics and Data Management

5. Budget

Provide an itemized list of anticipated costs (materials, personnel, etc.). These are not binding and are able to be adjusted based on the results of the pilot data.

6. References

List all references used in the manuscript using Nature style (e.g. numbered in-text citations and alphabetical references at the end of the manuscript).

1e:[["$","$L53",null,{"work":"$1d:props:children:props:topBanner:props:work","metadata":"$1d:props:children:props:topBanner:props:metadata","content":"$54","authorPosts":[]}],"$L55","$L56"]