Adding updated notes on FFPEsig and how to setup and install tools better

Author: wilkb777

Dockerfile

@@ -80,6 +80,55 @@ Following successful processing the wrapper script will have created a directory
 where the FFPE induced variants have been removed. These can be used for downstream variant and mutational signature
 identification.
 
+## Confirmation of FFPE induced variant removal using FFPEsig
+
+FFPEsig was developed to correct mutational signatures by adjusting for the expected signatures of FFPE induced
+variants. In this way we can use it as an orthogonal tool to Ideafix to confirm that FFPE induced variants were
+correctly identified and removed. Essentially if FFPEsig determines that little to no correction is needed in changing
+the mutational signature profile then it validates the removal process performed by Ideafix.
+
+### Setup
+
+First build the Anaconda3 environment for the tool
+
+```
+conda env create -f env/ffpesig-env.yml
+```
+
+or if you're using the Mamba wrapper/replacement for Conda
+
+```
+mamba env create -f env/ffpesig-env.yml
+```
+
+After the environment's been created activate it using
+
+```
+conda activate ffpesig
+```
+
+then download the FFPEsig Python script from the authors GitHub repo to run (this has already been done for this repo)
+
+```
+curl -L https://raw.githubusercontent.com/QingliGuo/FFPEsig/main/FFPEsig.py -o etc/ffpesig/FFPEsig.py
+```
+
+Now you're ready to run the tool!
+
+### Running FFPEsig
+
+With the FFPEsig script downloaded the tool has a CLI and can be run on the count matrix output by SigProfilerExtractor.
+
+```
+cd etc/ffpesig
+python FFPEsig.py -i data/ffpesig_input/ffpefixed_tumor_normal_called_all_hiconf.SBS96.all.csv -s "CHORANGIOMA" -l Unrepaired -o ../../data/ffpesig_output
+python FFPEsig.py -i data/ffpesig_input/ffpefixed_tumor_normal_called_all_hiconf.SBS96.all.csv -s "NORM_VILLI" -l Unrepaired -o ../../data/ffpesig_output
+```
+
+Results will appear in (data/ffpesig_output)[data/ffpesig_output] and comparing the `before_correction` to the
+`after_correction` results for each tissue will confirm that minimal correction was needed by FFPESig, confirming
+that Ideafix correctly removed spurious variants caused by the FFPE fixation process.
+
 ## Somatic mutation signal identification
 
 Mutational signal profiling from somatic variation to detect disrupted mechanisms in somatic cells (typically cancers)
@@ -174,92 +223,4 @@ Double, and indel signatures identified (both De Novo and Composite from the
 [COSMIC database of reference signatures](https://cancer.sanger.ac.uk/signatures/)). The outputs are in labeled
 subdirectories per signature type (SBS, DBS, IDS) and contains PDFs of signatures as well as signature stats and other
 accompanying information used in signature identification. More information can be found on the SigProfiler Wiki and
-Youtube videos.
-
-### Failed setup and use of SigProfiler
-
-This are simple notes marking past attempts for using SigProfiler which didn't work correctly. This is kept purely for
-reminder notes on what was done in the development of the wrapper for SigProfiler.
-
-#### Docker on MacOS
-
-This did not end well... It ran with max resources on all CPUs and did not complete. It seemingly stopped working after
-about 24 hours and could have been due to a number of factors...
-
-1. build the docker image for build 38 and the simple CLI tool for using it
-    ```
-    docker build -t sigprof-testing:v0.0.1 .
-    ```
-2. run the container specifying inner path to vcf file(s) directory that was mounted
-    ```
-    docker run --rm -v "$(pwd)"/data:/usr/sigprof/vcfs -v "$(pwd)"/output:/usr/sigprof/results -it --entrypoint bash sigprof-testing:v0.0.1
-    ```
-
-#### Conda environment on MacOS
-
-This setup ran but was hard coded to certain paths and very specific to a single machine setup. Ultimately the more this
-tool was needing to be run the more difficult it was to use this.
-
-1. build the conda env
-    1. `conda env create -f env/sigprofiler-env.yml`
-2. activate the env
-    1. `conda activate sigprofiler`
-3. install the reference genome materials needed to run the sigprofiler
-    1. `cd data`
-    2. `python -c "from SigProfilerMatrixGenerator import install as genInstall; genInstall.install('GRCh38')"`
-    3. _NOTE_: the profiler, or the script it's run from, must run from this directory as well due to it expecting a
-       certain path for this information being installed
-    4. _WARNING_: This can be a long running process so don't run in a setting where the task could be accidentally
-       terminated
-4. start an interactive python session
-    1. `python`
-5. Import SigProfileExtractor and run the application
-    1. ```
-           from SigProfilerExtractor import sigpro as sig
-           sig.sigProfilerExtractor(
-               input_type="vcf",
-               output="results",
-               input_data="pvp_input",
-               reference_genome="GRCh38",
-               opportunity_genome="GRCh38",
-               minimum_signatures=1,
-               maximum_signatures=10,
-               cosmic_version=3.3,
-               collapse_to_SBS96=False,
-               get_all_signature_matrices=False)
-       ```
-6. the profiler should complete relatively quickly with a single sample. Results will appear in `data/results/`
-
-## Somatic Driver variant identification
-
-Testing new tool reported in
-["Genome-wide mapping of somatic mutation rates uncovers drivers of cancer"](https://www.nature.com/articles/s41587-022-01353-8)
-that should aid in driver variant (previously identified and novel) identification including those in non-coding and
-regulatory regions that could be confirmed with RNASeq.
-
-**WARNING**: The authors of this tool note that use of the tool could be "spurious" when analyzing a type of cancer
-against a pretrained model that doesn't match the mechanisms/signatures of cancers provided in the pretrained models.
-see this [GitHub Issue](https://github.com/maxwellsh/DIGDriver/issues/2#issuecomment-1168689782) explaining more from
-the authors of the tool.
-
-**Annoyance**: The authors recommend using liftover to `hg19` perform analysis of datasets aligned to `GRCh38/hg38`. See
-their [recommendation](https://github.com/maxwellsh/DIGDriver/issues/2#issuecomment-1168689782) notes on this process.
-This isn't the greatest but if it's what it take to use the tool it may be an acceptable limitation.
-
-### Conda environment
-
--   [Conda environment info](https://anaconda.org/mutation_density/digdriver)
--   [Install and setup wiki](https://github.com/maxwellsh/DIGDriver/wiki)
-
-1. Build the conda environment as defined in the [yaml file](./env/digdriver-env.yml) for DigDriver
-    1. ```sh
-         conda env create -f env/digdriver-env.yml
-       ```
-2. Download the required [supporting data files](https://github.com/maxwellsh/DIGDriver/wiki/02:-Data-files) for running
-   DigDiver
-    1. Download the pretrained models that might work out of the box ( Brandon's best guess ¯\_(ツ)\_/¯ ) using the
-       [download script](./setup/downloadDigModels.sh)
-    2. ```sh
-         ./setup/downloadDigModels.sh
-       ```
-    3. All supporting data will be downloaded under `data/dig_driver/` for use with DigDriver
+Youtube videos.

README.md

@@ -0,0 +1,11 @@
+name: ffpesig
+
+channels:
+  - conda-forge
+
+dependencies:
+  - python==3.7.6
+  - pandas==1.0.1
+  - numpy==1.18.1
+  - matplotlib==3.1.3
+  - seaborn==0.10.1