Edit template and include analysis of peptigate run on human transcriptome

[taylorreiter]

PR checklist

• [x] Describe the changes you've made.
• [x] Describe any tests you have conducted to confirm that your changes behave as expected.
• [x] If you've added new software dependencies, make sure that those dependencies are included in the appropriate conda environments.
• [x] If you've added new functionality, make sure that the documentation is updated accordingly.

PR description

Stripping files out of the template

At first, I was going to orchestrate these tests with snakemake, but they're simple enough that I'm choosing to document them in READMEs in analysis-specific folders. I stripped out a lot of the extra template stuff because of that.

running peptigate on human transcriptome

the first test I did to see if peptigate works is to run it on the human transcriptome. This PR records the commands I ran to retrieve the human data, run peptigate, and has a notebook where I analyze the results.

I would love feedback from @borgesadair1 and @ecpierce about missing tests, unclear interpretations, or other things they'd like to see or thoughts they have. All of this is in the notebook.

Note I haven't updated the general documentation for this repo yet.

[review-notebook-app[bot]]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[review-notebook-app[bot]]

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:07Z ----------------------------------------------------------------

Typo: "times" of peptides to "types" of peptides

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:08Z ----------------------------------------------------------------

Typo, "note" to "not shown"

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:08Z ----------------------------------------------------------------

I'm not sure I understand the difference between this figure and the faceted one showing 30 or less vs 31 or more, I'm not sure what that distinction is supposed to be?

taylorreiter commented on 2024-03-22T16:52:45Z ----------------------------------------------------------------

added a comment describing this, thank you!

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:09Z ----------------------------------------------------------------

This was just a little confusing and I had to reread it a couple of times to understand the take-home point of these tables. I think it's because there's a bioactivity score within a category and then that score is summed up across different categories?

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:10Z ----------------------------------------------------------------

This is interesting that those with a BLAST hit for each of the three prediction tools have a range of bioactivity scores. I still can't tell fi this sum is the different types of bioactivity categories the peptide had a hit in or some quantitative score of how bioactive something is supposed to be?

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:11Z ----------------------------------------------------------------

Is this table here a repeat of the one above?

taylorreiter commented on 2024-03-22T17:06:29Z ----------------------------------------------------------------

yes, the one above is inline markdown, this is the code that generates the table. I can remove it in one of those locations if its too confusing

taylorreiter commented on 2024-03-22T17:34:59Z ----------------------------------------------------------------

update i rephrased the whole section above and took out the table, so I'm going to leave it here!

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:12Z ----------------------------------------------------------------

Ah ok no I think this one is per petpide the different categories of bioactivity that were hit? Maybe I'm still confused by what the above score is then.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:12Z ----------------------------------------------------------------

Might be a pain to go back and add this, but I think above you mentioned that for those that have a BLAST hit it's not quite 100%. Could you add a third category for BLAST hits that for your query peptide are almost identical to what is in the database? And the others are merely "close" hits? This may help distinguish what could be "biologically" real if those almost identical hits are still dispersed continuously like they are above

taylorreiter commented on 2024-03-22T18:00:53Z ----------------------------------------------------------------

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:13Z ----------------------------------------------------------------

Would these clustering results be more useful if you were trying to cluster these peptides compared to something else? What is the expected result if these were actually closely related or different sets of peptides vs not supposed to be peptides? Should they be clustered closer together or in different subgroups?

taylorreiter commented on 2024-03-22T18:01:51Z ----------------------------------------------------------------

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:14Z ----------------------------------------------------------------

Yeah something like this where maybe you have subsets that cluster together that might have the same bioactivity or function? Might be confirmatory bias here but the tSNE just seems to not be informative?

taylorreiter commented 7 months ago

added a comment describing this, thank you!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

yes, the one above is inline markdown, this is the code that generates the table. I can remove it in one of those locations if its too confusing

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

update i rephrased the whole section above and took out the table, so I'm going to leave it here!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Thanks @elizabethmcd, I think I addressed all of your comments with changes to the notebook. I especially cleaned up the explanation for what I was doing with the bioactivity data.

You may also note that all of the graphs have changed -- after doing some experiments with the spider mite transcriptome, I saw that we had a ton of sORF false positives. I added a new step in peptigate to try and filter these transcripts before even doing prediction, which has led to a decrease in sORF prediction in human. It isn't perfect, but it was necessary to reduce some noise.

• https://github.com/Arcadia-Science/peptigate/issues/24
• https://github.com/Arcadia-Science/peptigate/pull/26

• © Githubissues.
• Githubissues is a development platform for aggregating issues.

[review-notebook-app[bot]]

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:08Z ----------------------------------------------------------------

Typo, "note" to "not shown"

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:08Z ----------------------------------------------------------------

I'm not sure I understand the difference between this figure and the faceted one showing 30 or less vs 31 or more, I'm not sure what that distinction is supposed to be?

taylorreiter commented on 2024-03-22T16:52:45Z ----------------------------------------------------------------

added a comment describing this, thank you!

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:09Z ----------------------------------------------------------------

This was just a little confusing and I had to reread it a couple of times to understand the take-home point of these tables. I think it's because there's a bioactivity score within a category and then that score is summed up across different categories?

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:10Z ----------------------------------------------------------------

This is interesting that those with a BLAST hit for each of the three prediction tools have a range of bioactivity scores. I still can't tell fi this sum is the different types of bioactivity categories the peptide had a hit in or some quantitative score of how bioactive something is supposed to be?

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:11Z ----------------------------------------------------------------

Is this table here a repeat of the one above?

taylorreiter commented on 2024-03-22T17:06:29Z ----------------------------------------------------------------

yes, the one above is inline markdown, this is the code that generates the table. I can remove it in one of those locations if its too confusing

taylorreiter commented on 2024-03-22T17:34:59Z ----------------------------------------------------------------

update i rephrased the whole section above and took out the table, so I'm going to leave it here!

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:12Z ----------------------------------------------------------------

Ah ok no I think this one is per petpide the different categories of bioactivity that were hit? Maybe I'm still confused by what the above score is then.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:12Z ----------------------------------------------------------------

Might be a pain to go back and add this, but I think above you mentioned that for those that have a BLAST hit it's not quite 100%. Could you add a third category for BLAST hits that for your query peptide are almost identical to what is in the database? And the others are merely "close" hits? This may help distinguish what could be "biologically" real if those almost identical hits are still dispersed continuously like they are above

taylorreiter commented on 2024-03-22T18:00:53Z ----------------------------------------------------------------

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:13Z ----------------------------------------------------------------

Would these clustering results be more useful if you were trying to cluster these peptides compared to something else? What is the expected result if these were actually closely related or different sets of peptides vs not supposed to be peptides? Should they be clustered closer together or in different subgroups?

taylorreiter commented on 2024-03-22T18:01:51Z ----------------------------------------------------------------

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:14Z ----------------------------------------------------------------

Yeah something like this where maybe you have subsets that cluster together that might have the same bioactivity or function? Might be confirmatory bias here but the tSNE just seems to not be informative?

taylorreiter commented 7 months ago

added a comment describing this, thank you!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

yes, the one above is inline markdown, this is the code that generates the table. I can remove it in one of those locations if its too confusing

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

update i rephrased the whole section above and took out the table, so I'm going to leave it here!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Thanks @elizabethmcd, I think I addressed all of your comments with changes to the notebook. I especially cleaned up the explanation for what I was doing with the bioactivity data.

You may also note that all of the graphs have changed -- after doing some experiments with the spider mite transcriptome, I saw that we had a ton of sORF false positives. I added a new step in peptigate to try and filter these transcripts before even doing prediction, which has led to a decrease in sORF prediction in human. It isn't perfect, but it was necessary to reduce some noise.

• https://github.com/Arcadia-Science/peptigate/issues/24
• https://github.com/Arcadia-Science/peptigate/pull/26

• © Githubissues.
• Githubissues is a development platform for aggregating issues.

[review-notebook-app[bot]]

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:08Z ----------------------------------------------------------------

I'm not sure I understand the difference between this figure and the faceted one showing 30 or less vs 31 or more, I'm not sure what that distinction is supposed to be?

taylorreiter commented on 2024-03-22T16:52:45Z ----------------------------------------------------------------

added a comment describing this, thank you!

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:09Z ----------------------------------------------------------------

This was just a little confusing and I had to reread it a couple of times to understand the take-home point of these tables. I think it's because there's a bioactivity score within a category and then that score is summed up across different categories?

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:10Z ----------------------------------------------------------------

This is interesting that those with a BLAST hit for each of the three prediction tools have a range of bioactivity scores. I still can't tell fi this sum is the different types of bioactivity categories the peptide had a hit in or some quantitative score of how bioactive something is supposed to be?

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:11Z ----------------------------------------------------------------

Is this table here a repeat of the one above?

taylorreiter commented on 2024-03-22T17:06:29Z ----------------------------------------------------------------

yes, the one above is inline markdown, this is the code that generates the table. I can remove it in one of those locations if its too confusing

taylorreiter commented on 2024-03-22T17:34:59Z ----------------------------------------------------------------

update i rephrased the whole section above and took out the table, so I'm going to leave it here!

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:12Z ----------------------------------------------------------------

Ah ok no I think this one is per petpide the different categories of bioactivity that were hit? Maybe I'm still confused by what the above score is then.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:12Z ----------------------------------------------------------------

Might be a pain to go back and add this, but I think above you mentioned that for those that have a BLAST hit it's not quite 100%. Could you add a third category for BLAST hits that for your query peptide are almost identical to what is in the database? And the others are merely "close" hits? This may help distinguish what could be "biologically" real if those almost identical hits are still dispersed continuously like they are above

taylorreiter commented on 2024-03-22T18:00:53Z ----------------------------------------------------------------

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:13Z ----------------------------------------------------------------

Would these clustering results be more useful if you were trying to cluster these peptides compared to something else? What is the expected result if these were actually closely related or different sets of peptides vs not supposed to be peptides? Should they be clustered closer together or in different subgroups?

taylorreiter commented on 2024-03-22T18:01:51Z ----------------------------------------------------------------

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:14Z ----------------------------------------------------------------

Yeah something like this where maybe you have subsets that cluster together that might have the same bioactivity or function? Might be confirmatory bias here but the tSNE just seems to not be informative?

taylorreiter commented 7 months ago

added a comment describing this, thank you!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

yes, the one above is inline markdown, this is the code that generates the table. I can remove it in one of those locations if its too confusing

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

update i rephrased the whole section above and took out the table, so I'm going to leave it here!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Thanks @elizabethmcd, I think I addressed all of your comments with changes to the notebook. I especially cleaned up the explanation for what I was doing with the bioactivity data.

You may also note that all of the graphs have changed -- after doing some experiments with the spider mite transcriptome, I saw that we had a ton of sORF false positives. I added a new step in peptigate to try and filter these transcripts before even doing prediction, which has led to a decrease in sORF prediction in human. It isn't perfect, but it was necessary to reduce some noise.

• https://github.com/Arcadia-Science/peptigate/issues/24
• https://github.com/Arcadia-Science/peptigate/pull/26

• © Githubissues.
• Githubissues is a development platform for aggregating issues.

[review-notebook-app[bot]]

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:09Z ----------------------------------------------------------------

This was just a little confusing and I had to reread it a couple of times to understand the take-home point of these tables. I think it's because there's a bioactivity score within a category and then that score is summed up across different categories?

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:10Z ----------------------------------------------------------------

This is interesting that those with a BLAST hit for each of the three prediction tools have a range of bioactivity scores. I still can't tell fi this sum is the different types of bioactivity categories the peptide had a hit in or some quantitative score of how bioactive something is supposed to be?

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:11Z ----------------------------------------------------------------

Is this table here a repeat of the one above?

taylorreiter commented on 2024-03-22T17:06:29Z ----------------------------------------------------------------

yes, the one above is inline markdown, this is the code that generates the table. I can remove it in one of those locations if its too confusing

taylorreiter commented on 2024-03-22T17:34:59Z ----------------------------------------------------------------

update i rephrased the whole section above and took out the table, so I'm going to leave it here!

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:12Z ----------------------------------------------------------------

Ah ok no I think this one is per petpide the different categories of bioactivity that were hit? Maybe I'm still confused by what the above score is then.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:12Z ----------------------------------------------------------------

Might be a pain to go back and add this, but I think above you mentioned that for those that have a BLAST hit it's not quite 100%. Could you add a third category for BLAST hits that for your query peptide are almost identical to what is in the database? And the others are merely "close" hits? This may help distinguish what could be "biologically" real if those almost identical hits are still dispersed continuously like they are above

taylorreiter commented on 2024-03-22T18:00:53Z ----------------------------------------------------------------

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:13Z ----------------------------------------------------------------

Would these clustering results be more useful if you were trying to cluster these peptides compared to something else? What is the expected result if these were actually closely related or different sets of peptides vs not supposed to be peptides? Should they be clustered closer together or in different subgroups?

taylorreiter commented on 2024-03-22T18:01:51Z ----------------------------------------------------------------

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:14Z ----------------------------------------------------------------

Yeah something like this where maybe you have subsets that cluster together that might have the same bioactivity or function? Might be confirmatory bias here but the tSNE just seems to not be informative?

taylorreiter commented 7 months ago

added a comment describing this, thank you!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

yes, the one above is inline markdown, this is the code that generates the table. I can remove it in one of those locations if its too confusing

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

update i rephrased the whole section above and took out the table, so I'm going to leave it here!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Thanks @elizabethmcd, I think I addressed all of your comments with changes to the notebook. I especially cleaned up the explanation for what I was doing with the bioactivity data.

You may also note that all of the graphs have changed -- after doing some experiments with the spider mite transcriptome, I saw that we had a ton of sORF false positives. I added a new step in peptigate to try and filter these transcripts before even doing prediction, which has led to a decrease in sORF prediction in human. It isn't perfect, but it was necessary to reduce some noise.

• https://github.com/Arcadia-Science/peptigate/issues/24
• https://github.com/Arcadia-Science/peptigate/pull/26

• © Githubissues.
• Githubissues is a development platform for aggregating issues.

[review-notebook-app[bot]]

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:10Z ----------------------------------------------------------------

This is interesting that those with a BLAST hit for each of the three prediction tools have a range of bioactivity scores. I still can't tell fi this sum is the different types of bioactivity categories the peptide had a hit in or some quantitative score of how bioactive something is supposed to be?

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:11Z ----------------------------------------------------------------

Is this table here a repeat of the one above?

taylorreiter commented on 2024-03-22T17:06:29Z ----------------------------------------------------------------

yes, the one above is inline markdown, this is the code that generates the table. I can remove it in one of those locations if its too confusing

taylorreiter commented on 2024-03-22T17:34:59Z ----------------------------------------------------------------

update i rephrased the whole section above and took out the table, so I'm going to leave it here!

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:12Z ----------------------------------------------------------------

Ah ok no I think this one is per petpide the different categories of bioactivity that were hit? Maybe I'm still confused by what the above score is then.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:12Z ----------------------------------------------------------------

Might be a pain to go back and add this, but I think above you mentioned that for those that have a BLAST hit it's not quite 100%. Could you add a third category for BLAST hits that for your query peptide are almost identical to what is in the database? And the others are merely "close" hits? This may help distinguish what could be "biologically" real if those almost identical hits are still dispersed continuously like they are above

taylorreiter commented on 2024-03-22T18:00:53Z ----------------------------------------------------------------

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:13Z ----------------------------------------------------------------

Would these clustering results be more useful if you were trying to cluster these peptides compared to something else? What is the expected result if these were actually closely related or different sets of peptides vs not supposed to be peptides? Should they be clustered closer together or in different subgroups?

taylorreiter commented on 2024-03-22T18:01:51Z ----------------------------------------------------------------

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:14Z ----------------------------------------------------------------

Yeah something like this where maybe you have subsets that cluster together that might have the same bioactivity or function? Might be confirmatory bias here but the tSNE just seems to not be informative?

taylorreiter commented 7 months ago

added a comment describing this, thank you!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

yes, the one above is inline markdown, this is the code that generates the table. I can remove it in one of those locations if its too confusing

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

update i rephrased the whole section above and took out the table, so I'm going to leave it here!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Thanks @elizabethmcd, I think I addressed all of your comments with changes to the notebook. I especially cleaned up the explanation for what I was doing with the bioactivity data.

You may also note that all of the graphs have changed -- after doing some experiments with the spider mite transcriptome, I saw that we had a ton of sORF false positives. I added a new step in peptigate to try and filter these transcripts before even doing prediction, which has led to a decrease in sORF prediction in human. It isn't perfect, but it was necessary to reduce some noise.

• https://github.com/Arcadia-Science/peptigate/issues/24
• https://github.com/Arcadia-Science/peptigate/pull/26

• © Githubissues.
• Githubissues is a development platform for aggregating issues.

[review-notebook-app[bot]]

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:11Z ----------------------------------------------------------------

Is this table here a repeat of the one above?

taylorreiter commented on 2024-03-22T17:06:29Z ----------------------------------------------------------------

yes, the one above is inline markdown, this is the code that generates the table. I can remove it in one of those locations if its too confusing

taylorreiter commented on 2024-03-22T17:34:59Z ----------------------------------------------------------------

update i rephrased the whole section above and took out the table, so I'm going to leave it here!

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:12Z ----------------------------------------------------------------

Ah ok no I think this one is per petpide the different categories of bioactivity that were hit? Maybe I'm still confused by what the above score is then.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:12Z ----------------------------------------------------------------

Might be a pain to go back and add this, but I think above you mentioned that for those that have a BLAST hit it's not quite 100%. Could you add a third category for BLAST hits that for your query peptide are almost identical to what is in the database? And the others are merely "close" hits? This may help distinguish what could be "biologically" real if those almost identical hits are still dispersed continuously like they are above

taylorreiter commented on 2024-03-22T18:00:53Z ----------------------------------------------------------------

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:13Z ----------------------------------------------------------------

Would these clustering results be more useful if you were trying to cluster these peptides compared to something else? What is the expected result if these were actually closely related or different sets of peptides vs not supposed to be peptides? Should they be clustered closer together or in different subgroups?

taylorreiter commented on 2024-03-22T18:01:51Z ----------------------------------------------------------------

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:14Z ----------------------------------------------------------------

Yeah something like this where maybe you have subsets that cluster together that might have the same bioactivity or function? Might be confirmatory bias here but the tSNE just seems to not be informative?

taylorreiter commented 7 months ago

added a comment describing this, thank you!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

yes, the one above is inline markdown, this is the code that generates the table. I can remove it in one of those locations if its too confusing

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

update i rephrased the whole section above and took out the table, so I'm going to leave it here!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Thanks @elizabethmcd, I think I addressed all of your comments with changes to the notebook. I especially cleaned up the explanation for what I was doing with the bioactivity data.

You may also note that all of the graphs have changed -- after doing some experiments with the spider mite transcriptome, I saw that we had a ton of sORF false positives. I added a new step in peptigate to try and filter these transcripts before even doing prediction, which has led to a decrease in sORF prediction in human. It isn't perfect, but it was necessary to reduce some noise.

• https://github.com/Arcadia-Science/peptigate/issues/24
• https://github.com/Arcadia-Science/peptigate/pull/26

• © Githubissues.
• Githubissues is a development platform for aggregating issues.

[review-notebook-app[bot]]

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:12Z ----------------------------------------------------------------

Ah ok no I think this one is per petpide the different categories of bioactivity that were hit? Maybe I'm still confused by what the above score is then.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:12Z ----------------------------------------------------------------

Might be a pain to go back and add this, but I think above you mentioned that for those that have a BLAST hit it's not quite 100%. Could you add a third category for BLAST hits that for your query peptide are almost identical to what is in the database? And the others are merely "close" hits? This may help distinguish what could be "biologically" real if those almost identical hits are still dispersed continuously like they are above

taylorreiter commented on 2024-03-22T18:00:53Z ----------------------------------------------------------------

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:13Z ----------------------------------------------------------------

Would these clustering results be more useful if you were trying to cluster these peptides compared to something else? What is the expected result if these were actually closely related or different sets of peptides vs not supposed to be peptides? Should they be clustered closer together or in different subgroups?

taylorreiter commented on 2024-03-22T18:01:51Z ----------------------------------------------------------------

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:14Z ----------------------------------------------------------------

Yeah something like this where maybe you have subsets that cluster together that might have the same bioactivity or function? Might be confirmatory bias here but the tSNE just seems to not be informative?

taylorreiter commented 7 months ago

added a comment describing this, thank you!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

yes, the one above is inline markdown, this is the code that generates the table. I can remove it in one of those locations if its too confusing

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

update i rephrased the whole section above and took out the table, so I'm going to leave it here!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Thanks @elizabethmcd, I think I addressed all of your comments with changes to the notebook. I especially cleaned up the explanation for what I was doing with the bioactivity data.

You may also note that all of the graphs have changed -- after doing some experiments with the spider mite transcriptome, I saw that we had a ton of sORF false positives. I added a new step in peptigate to try and filter these transcripts before even doing prediction, which has led to a decrease in sORF prediction in human. It isn't perfect, but it was necessary to reduce some noise.

• https://github.com/Arcadia-Science/peptigate/issues/24
• https://github.com/Arcadia-Science/peptigate/pull/26

• © Githubissues.
• Githubissues is a development platform for aggregating issues.

[review-notebook-app[bot]]

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:12Z ----------------------------------------------------------------

Might be a pain to go back and add this, but I think above you mentioned that for those that have a BLAST hit it's not quite 100%. Could you add a third category for BLAST hits that for your query peptide are almost identical to what is in the database? And the others are merely "close" hits? This may help distinguish what could be "biologically" real if those almost identical hits are still dispersed continuously like they are above

taylorreiter commented on 2024-03-22T18:00:53Z ----------------------------------------------------------------

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:13Z ----------------------------------------------------------------

Would these clustering results be more useful if you were trying to cluster these peptides compared to something else? What is the expected result if these were actually closely related or different sets of peptides vs not supposed to be peptides? Should they be clustered closer together or in different subgroups?

taylorreiter commented on 2024-03-22T18:01:51Z ----------------------------------------------------------------

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:14Z ----------------------------------------------------------------

Yeah something like this where maybe you have subsets that cluster together that might have the same bioactivity or function? Might be confirmatory bias here but the tSNE just seems to not be informative?

taylorreiter commented 7 months ago

added a comment describing this, thank you!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

yes, the one above is inline markdown, this is the code that generates the table. I can remove it in one of those locations if its too confusing

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

update i rephrased the whole section above and took out the table, so I'm going to leave it here!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Thanks @elizabethmcd, I think I addressed all of your comments with changes to the notebook. I especially cleaned up the explanation for what I was doing with the bioactivity data.

You may also note that all of the graphs have changed -- after doing some experiments with the spider mite transcriptome, I saw that we had a ton of sORF false positives. I added a new step in peptigate to try and filter these transcripts before even doing prediction, which has led to a decrease in sORF prediction in human. It isn't perfect, but it was necessary to reduce some noise.

• https://github.com/Arcadia-Science/peptigate/issues/24
• https://github.com/Arcadia-Science/peptigate/pull/26

• © Githubissues.
• Githubissues is a development platform for aggregating issues.

[review-notebook-app[bot]]

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:13Z ----------------------------------------------------------------

Would these clustering results be more useful if you were trying to cluster these peptides compared to something else? What is the expected result if these were actually closely related or different sets of peptides vs not supposed to be peptides? Should they be clustered closer together or in different subgroups?

taylorreiter commented on 2024-03-22T18:01:51Z ----------------------------------------------------------------

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

review-notebook-app[bot] commented 7 months ago

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:14Z ----------------------------------------------------------------

Yeah something like this where maybe you have subsets that cluster together that might have the same bioactivity or function? Might be confirmatory bias here but the tSNE just seems to not be informative?

taylorreiter commented 7 months ago

added a comment describing this, thank you!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

yes, the one above is inline markdown, this is the code that generates the table. I can remove it in one of those locations if its too confusing

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

update i rephrased the whole section above and took out the table, so I'm going to leave it here!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Thanks @elizabethmcd, I think I addressed all of your comments with changes to the notebook. I especially cleaned up the explanation for what I was doing with the bioactivity data.

You may also note that all of the graphs have changed -- after doing some experiments with the spider mite transcriptome, I saw that we had a ton of sORF false positives. I added a new step in peptigate to try and filter these transcripts before even doing prediction, which has led to a decrease in sORF prediction in human. It isn't perfect, but it was necessary to reduce some noise.

• https://github.com/Arcadia-Science/peptigate/issues/24
• https://github.com/Arcadia-Science/peptigate/pull/26

• © Githubissues.
• Githubissues is a development platform for aggregating issues.

[review-notebook-app[bot]]

View / edit / reply to this conversation on ReviewNB

elizabethmcd commented on 2024-03-20T19:31:14Z ----------------------------------------------------------------

Yeah something like this where maybe you have subsets that cluster together that might have the same bioactivity or function? Might be confirmatory bias here but the tSNE just seems to not be informative?

taylorreiter commented 7 months ago

added a comment describing this, thank you!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

yes, the one above is inline markdown, this is the code that generates the table. I can remove it in one of those locations if its too confusing

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

update i rephrased the whole section above and took out the table, so I'm going to leave it here!

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

---

View entire conversation on ReviewNB

taylorreiter commented 7 months ago

Thanks @elizabethmcd, I think I addressed all of your comments with changes to the notebook. I especially cleaned up the explanation for what I was doing with the bioactivity data.

You may also note that all of the graphs have changed -- after doing some experiments with the spider mite transcriptome, I saw that we had a ton of sORF false positives. I added a new step in peptigate to try and filter these transcripts before even doing prediction, which has led to a decrease in sORF prediction in human. It isn't perfect, but it was necessary to reduce some noise.

• https://github.com/Arcadia-Science/peptigate/issues/24
• https://github.com/Arcadia-Science/peptigate/pull/26

• © Githubissues.
• Githubissues is a development platform for aggregating issues.

[taylorreiter]

added a comment describing this, thank you!

---

View entire conversation on ReviewNB

[taylorreiter]

yes, the one above is inline markdown, this is the code that generates the table. I can remove it in one of those locations if its too confusing

---

View entire conversation on ReviewNB

[taylorreiter]

update i rephrased the whole section above and took out the table, so I'm going to leave it here!

---

View entire conversation on ReviewNB

[taylorreiter]

This is a really great idea. I tried doing this and the identical hits are still dispersed continuously -- it doesn't look much different than the plot here and the code is super messy so I'm not going to make the change if that's ok.

---

View entire conversation on ReviewNB

[taylorreiter]

Emily had this idea as well! like what if we tried to cluster the entire peptipedia database, and then overlaid our data with that. I'm going to leave these clusters as is for this notebook, but I totally think you guys are on to something and I'm going to keep thinking about the best way to do this.

---

View entire conversation on ReviewNB

[taylorreiter]

Thanks @elizabethmcd, I think I addressed all of your comments with changes to the notebook. I especially cleaned up the explanation for what I was doing with the bioactivity data.

You may also note that all of the graphs have changed -- after doing some experiments with the spider mite transcriptome, I saw that we had a ton of sORF false positives. I added a new step in peptigate to try and filter these transcripts before even doing prediction, which has led to a decrease in sORF prediction in human. It isn't perfect, but it was necessary to reduce some noise.

• https://github.com/Arcadia-Science/peptigate/issues/24
• https://github.com/Arcadia-Science/peptigate/pull/26