Overview

Welcome to the GitHub repository for the following publication: The mutational landscape of a prion-like domain (Bolognesi B & Faure AJ et al., 2019)

Here you'll find an R package with all scripts to reproduce the figures and results from the computational analyses described in the paper.

Required Software

To run the tardbpdms pipeline you will need the following software and associated packages:

• R >=v3.5.2 (Biostrings, caTools, corpcor, cowplot, data.table, gdata, ggplot2, GGally, hexbin, lemon, optparse, parallel, pdist, plyr, ppcor, raster, reshape2, Rpdb, RColorBrewer)

The following packages are optional:

• DiMSum (pipeline for pre-processing deep mutational scanning data i.e. FASTQ to counts)
• DMS2structure (scripts used for epistasis and structure analysis of deep mutational scanning data in Schmiedel & Lehner, bioRxiv 2018)

Installation and loading

Open R and enter:

# Install
if(!require(devtools)) install.packages("devtools")
devtools::install_github("lehner-lab/tardbpdms")

# Load
library(tardbpdms)

# Help
?tardbpdms

Required Data

Variant counts, pre-processed data and required miscellaneous files should be downloaded from here to your project directory (see 'base_dir' argument) i.e. where output files should be written, and unzipped.

Running

There are a number of options available for running the tardbpdms pipeline depending on user requirements.

• Basic (default)

Default pipeline functionality uses variant counts (see 'Required Data') to reproduce all figures in the publication. Neither DiMSum nor DMS2structure packages are required for this default functionality.

• Raw read processing

Raw read processing is not handled by the tardbpdms pipeline. FastQ files (GSE128165) from paired-end sequencing of replicate deep mutational scanning (DMS) libraries before ('input') and after selection ('output') were processed using DiMSum (manuscript in prep.), an R package that wraps common biological sequence processing tools.

DiMSum command-line arguments and Experimental design files required to obtain variant counts from FastQ files are available here.

• Estimating fitness (1-toxicity) from variant counts

Pipeline stage 1 ('tardbpdms_dimsumcounts_to_fitness') estimates toxicity and error of single and double AA mutants from variant counts for each library separately. This stage is computationally intensive (~2hours on 10 cores) and is therefore not run by default. Note: When running the pipeline for the first time or to force re-execution of this stage set 'rerun_fitness = T'.

• Epistasis analysis

Pipeline stage 11 ('tardbpdms_epistasis_analysis') performs epistasis calculations separately for each DMS library. This stage is computationally intensive (~1hour on 10 cores) and is therefore not run by default. Note: 'Required Data' (see above) already includes precomputed results of the epistasis analysis necessary to reproduce the corresponding figures in the publication. However, to force re-execution of this stage set 'rerun_epistasis = T'. Additionally, the correct path to your local copy of the DMS2structure repository must be specified with 'DMS2structure_path = MY_LOCAL_PATH'.

• Structure analyses

Pipeline stages 12 and 13 ('tardbpdms_secondary_structure_predictions', 'tardbpdms_guenther_structure_propensities') perform secondary structure predictions and structure propensity calculations for PDB-structure derived contact matrices respectively. Secondary structure predictions and propensity calculations are computationally intensive and are therefore not re-run by default. Note: 'Required Data' (see above) already includes precomputed results of the structure analyses necessary to reproduce the corresponding figures in the publication. To force re-execution set 'rerun_structure = T'. Additionally, the correct path to your local copy of the DMS2structure repository must be specified with 'DMS2structure_path = MY_LOCAL_PATH'.

Pipeline

The top-level function tardbpdms() is the recommended entry point to the pipeline and reproduces the figures and results from the computational analyses described in the following publication: "The mutational landscape of a prion-like domain" (Bolognesi B & Faure AJ et al., 2019). See section on "Required Data" above for instructions on how to obtain all required data and miscellaneous files before running the pipeline.

Stage 1: DiMSum counts to fitness

This stage ('tardbpdms_dimsumcounts_to_fitness') estimates toxicity and error of single and double AA mutants from variant counts for each library separately. This stage is computationally intensive (~2hours on 10 cores) and is therefore not run by default. When running the pipeline for the first time or to force re-execution of this stage set 'rerun_fitness = T'.

Stage 2: Quality control plots

This stage ('tardbpdms_quality_control') produces quality control plots of toxicity estimates before and after inter-replicate normalisation.

Stage 3: Combine toxicity estimates from 290 and 332 experiments

This stage ('tardbpdms_combine_toxicity') performs inter-library normalisation, toxicity distribution plots, growth rate comparison plots and position-wise toxicity plots.

Stage 4: Calculate single and double mutant effects from AA PCA

This stage ('tardbpdms_aa_properties_mutant_effects') performs principal component analysis (PCA) of a curated collection of numerical indices representing various physicochemical and biochemical properties of amino acid (AA) properties. AA property feature values represent the difference between the WT and mutant PC scores.

Stage 5: Calculate single and double mutant effects from aggregation tool predictions

This stage ('tardbpdms_agg_tools_mutant_effects') calculates aggregation / disorder algorithm feature values for single and double mutant variants (similar to stage 4).

Stage 6: Single mutant heatmaps

This stage ('tardbpdms_single_mutant_heatmaps') produces single mutant heatmaps of toxicity effects.

Stage 7: Human disease mutations

This stage ('tardbpdms_human_disease_mutations') tests whether human disease mutations have biased toxicity estimates.

Stage 8: Dot plots showing explained variance of models to predict variant toxicity

This stage ('tardbpdms_toxicity_model_summary') produces plots of results from simple linear regression models to predict variant toxicity.

Stage 9: Violin plots showing toxicity vs #introduced AAs

This stage ('tardbpdms_num_introduced_aa_violins') produces violin plots and scatterplots of toxicity (distributions) versus hydrophobicity, aromaticity, charge and aggregation propensity.

Stage 10: Hydrophobicity of WT and toxicity hotspot line plot

This stage ('tardbpdms_wt_hydrophobicity') plots hydrophobicity score and mean toxicity effect along the length of WT TDP-43.

Stage 11: Epistasis analysis

This stage ('tardbpdms_epistasis_analysis') performs epistasis calculations separately for each DMS library. This stage is computationally intensive (~1hour on 10 cores) and is therefore not run by default. To force re-execution of this stage set 'rerun_epistasis = T'. Additionally, the corect path to your local copy of the DMS2structure repository must be specified with 'DMS2structure_path = MY_LOCAL_PATH'.

Stage 12: Secondary structure predictions

This stage ('tardbpdms_secondary_structure_predictions') performs secondary structure predictions for each DMS library and produces combined summary plots. Secondary structure predictions are computationally intensive and are therefore not re-run by default. To force re-execution of secondary structure predictions set 'rerun_structure = T'. Additionally, the corect path to your local copy of the DMS2structure repository must be specified with 'DMS2structure_path = MY_LOCAL_PATH'.

Stage 13: Guenther structure propensities

This stage ('tardbpdms_guenther_structure_propensities') performs structure propensity calculations for PDB-structure derived contact matrices. Structure propensity calculation are computationally intensive and are therefore not re-run by default. To force re-execution of structure propensity calculations set 'rerun_structure = T'. Additionally, the corect path to your local copy of the DMS2structure repository must be specified with 'DMS2structure_path = MY_LOCAL_PATH'.

Stage 14: PWI heatmaps

This stage ('tardbpdms_PWI_heatmaps') plots pair-wise interaction (PWI) score heatmaps.