Add the two-pass search tutorial.

Nesvilab · Dec 17, 2023 · 349fbfe · 349fbfe
1 parent 27017f7
commit 349fbfe
Show file tree

Hide file tree

Showing 2 changed files with 86 additions and 0 deletions.
diff --git a/README.md b/README.md
@@ -29,6 +29,7 @@ FragPipe is a Java Graphical User Interface (GUI) for a suite of computational t
 * [Label-free quantification](https://fragpipe.nesvilab.org/docs/tutorial_lfq.html)
 * [SILAC (or other MS1-labeled) data](https://fragpipe.nesvilab.org/docs/tutorial_silac.html)
 * [DIA analysis](https://fragpipe.nesvilab.org/docs/tutorial_DIA.html)
+* [Novel/variant peptide detection](https://fragpipe.nesvilab.org/docs/tutorial_two_pass_search.html)
 
 #### Resources
 * [Interpreting output files](https://fragpipe.nesvilab.org/docs/tutorial_fragpipe_outputs.html)

diff --git a/docs/tutorial_two_pass_search.md b/docs/tutorial_two_pass_search.md
@@ -0,0 +1,85 @@
+## Perform two-pass search to analyze variant peptides
+
+##### [FragPipe](https://fragpipe.nesvilab.org) can be downloaded [here](https://github.com/Nesvilab/FragPipe/releases). Follow the instructions on that same Releases page to launch the program. See [here](https://fragpipe.nesvilab.org/docs/tutorial_fragpipe.html#configure-fragpipe) for help configuring FragPipe.
+
+FragPipe has the ability to perform the "two-pass search" to detect novel/variant peptides semi-automatically. It performs the following three steps.
+1. Perform the first search to identify the canonical peptides
+2. Generate new "sub mzML" files by excluding the scans matching to the canonical peptides
+3. Perform the second search to identify the novel/variant peptides
+
+#### Common issues of two-pass search approaches due to the limited high-quality scans in the second search
+1. The spectra are not mass calibrated or the mass calibration does not work well
+2. The rescoring tools, such as Percolator, cannot learn a good model in the second search
+
+#### To address these issues, FragPipe 
+1. performs mass calibration in the first search, and then generates sub mzML files using the calibrated spectra
+2. writes the Percolator models (`*_percolator.weights` files) to the result directory after the first search. These models will be used in the second search.
+
+The two-pass search approach supports the `Default`, `LFQ-MBR`, and TMT-related workflows.
+
+### Perform two-pass search for the label-free quantification data
+1. Load the `LFQ-MBR` workflow in the `Workflow` tab
+2. Load the LC-MS files, assign experiments and bioreplicates (if applicable), and set the `Data type` to `DDA`
+3. Load your FASTA file in the `Database` tab
+4. Go to the `Run` tab, specify the output directory
+5. **[Critical]** In the `Run` tab, enable the `Write sub mzML`. FragPipe will generate "sub mzML" files containing the scans do not pass the FDR filtering and have the probability smaller than or equal to the `Min probability threshold of excluding scans from sub mzML files`. 
+
+![](https://raw.githubusercontent.com/Nesvilab/FragPipe/gh-pages/images/two_pass_search_1.png)
+
+6. Click run. FragPipe will perform the "first-pass search".
+7. After FragPipe finish, go back to the `Workflow` tab
+8. Select the `Custom` workflow and click `Load workflow`
+
+![](https://raw.githubusercontent.com/Nesvilab/FragPipe/gh-pages/images/two_pass_search_2.png)
+
+9. Locate the `fragpipe-second-pass.workflow` file in the result directory and load it
+10. Click `Clear files` to remove all LC-MS files used in the first-pass search
+11. Click `Load manifest`, locate the `fragpipe-second-pass.manifest` file in the result directory, and load it
+
+![](https://raw.githubusercontent.com/Nesvilab/FragPipe/gh-pages/images/two_pass_search_3.png)
+
+12. Go to the `Database` tab to specify the FASTA file for the second search.
+13. Go to `Run` tab, specify the new output directory and click `Run`
+
+
+### Perform two-pass search for the isobaric-labeling data
+1. Load one of the TMT workflow in the `Workflow` tab
+2. Load the LC-MS files, assign experiments and bioreplicates (if applicable), and set the `Data type` to `DDA`
+3. Load your FASTA file in the `Database` tab
+4. Go to the `Quant (Isobaric)` tab, configure the sample/channel annotations. Details can be found in the [other tutorial](https://fragpipe.nesvilab.org/docs/tutorial_tmt-2plexes.html)
+5. Go to the `Run` tab, specify the output directory
+6. **[Critical]** In the `Run` tab, enable the `Write sub mzML`. FragPipe will generate "sub mzML" files containing the scans do not pass the FDR filtering and have the probability smaller than or equal to the `Min probability threshold of excluding scans from sub mzML files`. 
+
+![](https://raw.githubusercontent.com/Nesvilab/FragPipe/gh-pages/images/two_pass_search_1.png)
+
+7. Click run. FragPipe will perform the "first-pass search".
+8. After FragPipe finish, go back to the `Workflow` tab
+9. Select the `Custom` workflow and click `Load workflow`
+
+![](https://raw.githubusercontent.com/Nesvilab/FragPipe/gh-pages/images/two_pass_search_2.png)
+
+10. Locate the `fragpipe-second-pass.workflow` file in the result directory and load it
+11. Click `Clear files` to remove all LC-MS files used in the first-pass search
+12. Click `Load manifest`, locate the `fragpipe-second-pass.manifest` file in the result directory, and load it
+
+![](https://raw.githubusercontent.com/Nesvilab/FragPipe/gh-pages/images/two_pass_search_3.png)
+
+13. Go to the `Database` tab to specify the FASTA file for the second search.
+14. Go to the `Quant (Isobaric)` tab to ensure that the sample/channel annotation is correct.
+15. Go to `Run` tab, specify the new output directory and click `Run`
+
+
+
+<br>
+<br>
+
+
+#### Key References
+Desai H, Ofori S, Boatner L, Yu F, Villanueva M, Ung N, Nesvizhskii AI, Backus K. [Multi-omic stratification of the missense variant cysteinome](https://doi.org/10.1101/2023.08.12.553095), bioRxiv (2023).
+
+
+
+<br>
+<br>
+
+#### [Back to FragPipe homepage](https://fragpipe.nesvilab.org/)