semiller10 / postnovo

Post-processing peptide de novo sequences to improve their accuracy

Post-processing peptide de novo sequences to improve their accuracy

Journal article: https://pubs.acs.org/doi/10.1021/acs.jproteome.8b00278

Quick Start

Much more detail on Postnovo options can be found in the Postnovo Wiki pages.

1. Postnovo runs on a Unix-like OS. Use a powerful server or desktop with >25 GB available disk space.

2. Download and decompress the latest release.

3. Use Python 3. The Anaconda distribution contains all necessary package dependencies.

4. Download DeNovoGUI and large pre-trained models with Postnovo setup.

   For low-res MS2 data (example uses nohup to avoid termination upon logout and & to run in background):

   nohup python main.py setup --denovogui --postnovo_low --deepnovo_low &

   For high-res MS2 data:

   python main.py setup --denovogui --postnovo_high --deepnovo_high

5. Use the ProteoWizard msconvert tool to convert your RAW file to an MGF file with a certain spectrum header format.

   msconvert preformatted_spectra.raw --mgf --filter "titleMaker Run: <RunId>, Index: <Index>, Scan: <ScanNumber>"

   Add additional filters as needed, here for peak picking and removal of zero intensity peaks.

   msconvert preformatted_spectra.raw --mgf --filter "peakPicking vendor" --filter "zeroSamples removeExtra" --filter "titleMaker Run: <RunId>, Index: <Index>, Scan: <ScanNumber>"

6. Reformat the input MGF file for compatability with all de novo sequencing tools.

   python main.py format_mgf --mgfs /path/to/preformatted_spectra.mgf --out /path/to/spectra.mgf

7. Set up a container with TensorFlow to run DeepNovo (optional but recommended for Postnovo). Superuser privileges may be required.

   singularity build tensorflow.simg docker://tensorflow/tensorflow:latest

8. Generate DeepNovo de novo sequences (can take up to ~12 hours depending on resources). The following examples consider low-res MS2 spectra. Processing high-res spectra requires more memory (see Predicting with Deepnovo). Postnovo only supports standard fixed C and variable M PTMs at the moment.

   Using a single machine with 32 cores:

   python main.py predict_deepnovo --mgf /path/to/spectra.mgf --container /path/to/tensorflow.simg --frag_resolution low --cpus 32

   Using a compute cluster via Slurm with 16 cores per node and sufficient memory allocation:

   python main.py predict_deepnovo --mgf /path/to/spectra.mgf --container /path/to/tensorflow.simg --frag_resolution low --cpus 16 --slurm --partition partition_with_16GB_mem --time_limit 36

9. Download MaRaCluster to cluster spectra by peptide species. As MaRaCluster input, use the reformatted MGF file. Set "log10(p-value) threshold" (in the GUI) or "--pvalThreshold" (in the CLI) to -2.

10. Run Postnovo, which in the process generates Novor and PepNovo+ de novo sequences via DeNovoGUI (can take up to ~3 hours depending on resources). Results are written by default to the directory of the MGF input.

    python main.py predict --mgf /path/to/spectra.mgf --clusters /path/to/MaRaCluster.clusters_p2.tsv --frag_method CID --frag_resolution low --denovogui --deepnovo --cpus 32

Copyright 2018, Samuel E. Miller. All rights reserved.

Postnovo is publicly available for non-commercial uses.

Licensed under GNU GENERAL PUBLIC LICENSE, Version 3, 29 June 2007.

See postnovo/LICENSE.txt.

postnovo's Issues

setup error

Hi,

I would like to test postnovo but I got an error message while I try to set up it. Actually, the same error for all command lines, e.g, python main.py setup --deepnovo_low --deepnovo_high --container /path/container or ´python main.py setup --check_updates`

> python3 main.py setup --deepnovo_low --deepnovo_high --container /mnt/DATA_DISKS/WORKSPACE2/cguetot/installers/postnovo/tensorflow.simg
Traceback (most recent call last):
File "main.py", line 12, in
import interspec
File "/mnt/DATA_DISKS/WORKSPACE2/cguetot/installers/postnovo/postnovo/interspec.py", line 100
global cluster_dfs
^
SyntaxError: name 'cluster_dfs' is used prior to global declaration

or

> python3 main.py setup --check_update
Traceback (most recent call last):
File "main.py", line 12, in
import interspec
File "/mnt/DATA_DISKS/WORKSPACE2/cguetot/installers/postnovo/postnovo/interspec.py", line 100
global cluster_dfs
^
SyntaxError: name 'cluster_dfs' is used prior to global declaration

I tried with version 1.0.9-alpha and also from the master branch.

thank you in advance,

Carlos

Not all Google drive locations accessible

Dear Samuel, Thanks for the nice tool, I've been trying to get it to work, but I am running into problems with downloading your files from the google drive.
I ended up trying gdown as well, but some files have permission problems, and do not appear when searched with URL either.

Common error message:
Access denied with the following error:

Cannot retrieve the public link of the file. You may need to change
the permission to 'Anyone with the link', or have had many accesses. 

You may still be able to access the file from the browser:

 https://drive.google.com/uc?id=1XAipgL7afq16rHgJ_tOsAW7n7_YFVOc_ 

But the supplied link also does not seem to work.

Do you have any suggestions?
Kind regards,
Hugo Kleikamp

List of files that work or not:

<style> </style>

	Download Filename	Google Drive ID	Size	Access (yes/no)
0	DeNovoGUI-1.16.2-mac_and_linux.tar.gz	12WKUfRMpNjY7YilRg7zGNISgycNBsZ9M	77725399	y
1	MSGFPlus.zip	1k66SctfqBQyYAdKhQJbie005-ZDPfZyb	19202332	y
2	postnovo_low_default_models.zip	1k-jSc4SG-yqJf2vYXoEE_dk0xYTHMaBK	882282390	y
3	postnovo_high_default_models.zip	1xkxT5_oVydFRMzwMcGIjncKSrlSHEPPR	882474786	y
4	postnovo_low_default_training_data.zip	1iIxJB9JdPg-eDI7v_LLYSrpzKK_nHRJm	620061720	n
5	postnovo_high_default_training_data.zip	1JvoWm4mtWDGO1RKengTZiIrtas2tC-E1	709973469	n
6	train_record_low.tsv	1pd4k9ETr8UutNIfioC7zvfAh5Rx4rG8m	244	y
7	train_record_high.tsv	1IennZj1QxxUOyDeIZZw3Z2ohuWneXg5U	586	y
8	deepnovo_low_default_models.zip	1Fx_gCoXzUTr7FMgmYpbL5xYkuY8-nR6-	820270761	y
9	deepnovo_high_default_models.zip	1v3F8ZKU88Gz7ViEWDlUVt-W6orYBLdmc	3192686775	y
10	knapsack.npy	1N_CQPQlSbAW_fShocE9gMB1oFUjcZKOY	775577350	y
11	postnovo_low_default_spectra.zip	1H0yD5jhrcvAzxNf1fIr6SSoR9wDUKnlB	2521313775	y
12	postnovo_high_default_spectra.zip	1XAipgL7afq16rHgJ_tOsAW7n7_YFVOc_	1507107028	n
13	deepnovo_low_default_spectra.zip	1Bbz0QAqmAgizb6sbVDWNV791AA9fjx9U	521178064	y
14	deepnovo_high_default_spectra.zip	1JkgBJM6qsE8BB2kgBkrl8JprO2fHO2Oj	299785319	y

problems with tensorflow and python version

Hello,
I encountered problems with tensorflow version (2.3.1) contained in the singularity container. Currently, module are not found when called for import in each python deepnovo script. suggestions are to replace "import tensorflow as tf" by "import tensorflow.compat.v1 as tf".
Do you have any ideas or suggestions about it ?
thank you in advance,
benjamin Dartigues