Despres_et_al_2021

Notebooks for data analysis and figure generation of Despres et al 2021, "Asymmetrical dose-responses shape the evolutionary trade-off between antifungal resistance and nutrient use"

The notebooks for the Deep Mutationnal Scanning (DMS) data analysis and growth curves are listed here. The main analysis steps for sequencing data are also explained here (as well as in the methods section of the manuscript). Most bioinformatics was performed in Jupyter Notebooks using Python 3.7.9.

Python Packages

• pandas 1.2.1
• numpy 1.19.2
• matplotlib 3.3.2
• scipy 1.5.2
• seaborn 0.11.1
• DSSParser 0.10

Programs required:

• FastQC 0.11.5
• Trimmomatic 0.39 (make sure to change the path in the notebooks to the one appropriate for your environment)
• pandaseq 2.11
• vsearch 2.7.1
• needle (as part of EMBOSS 6.6.0.0)

DMS Notebooks included in the repository:

• DMS_analysis: details the pipeline to go from the MiSEQ fastq file to relative abudances of each FCY1 codon level variant at each timepoints.
• Fold_change: details the data analysis steps to go from codon level relative abundance to amino acid level log2 fold-change.
• Pymol_analysis: integrates the FCY1 FoldX data and other protein level information from various database in to the main DataFrame.
• Heatmaps: main notebook used to generate figures and perform statistical analysis.

Growth curve analysis notebooks included in the repository:

• validations: Analyze growth curve results and DHFR-PCA experiments
• hill_coefficient: analyze dose-response growth curves
• C_neo_mutants: Generate oligonucleotides for cnFCY1 mutagenesis and analyze experiments results

DMS_analysis

This notebook processes the fastq files and performs the following operations:

1. Assess reads quality using FastQC
2. Select reads with correct length and crop the 3-prime end to remove low confidence bases
3. Demultiplex the libraries that were combined using the plate RC-PCR library construction workflow

Then, for each library

4. Merge R1 and R2 using Pandaseq
5. Trim amplicons to remove adapter and barcode regions
6. Aggregate identical reads together
7. Align reads on the appropriate FCY1 cds reference
8. Parse alignment to detect mutations
9. Filter alignments and determine the effects of the mutations on the FCY1 cds at the codon level
10. Sum variant abundances and calulate relative abundance of each FCY1 cds variant, and save as csv for each library. we also add 1 to all read counts to make log2 fold-change calculations possible in cases where variants dropped out of the pool during the competition

We applied stringent filters to minimize noise due to sequencing errors. At step 7, all sequences that have > 25 differences with the reference or that cover less than 80% of it are dropped. Most importantly, we only considered reads where mutations in the coding sequence occured within the same codon to avoid including double mutants.