Calculating site-specific evolutionary rates at the amino-acid or codon level yields similar rate estimates

This repository contains all the scripts and data to reproduce the results of:

D. K. Sydykova, C. O. Wilke (2017). Calculating site-specific evolutionary rates at the amino-acid or codon level yields similar rate estimates. PeerJ 5:e3391. https://doi.org/10.7717/peerj.3391

Contents

mech_codon contains results for the alignments simulated with the dN/dS model.

• assigned_rates contains true site-wise dN/dS.

• filtered_sites contains information on all sites without any amino acid substitutions for each simulated alignment.

• inferred_rates contains inferred site-wise dN/dS.

• processed_rates contains tables with all site-wise rates: true dN/dS, inferred dN/dS, and inferred Rate4Site.

• r4s_rates contains inferred site-wise Rate4site rates.

mut_sel contains results for the alignments simulated with the mutation-selection (MutSel) model. MutSel alignments were simulated by Spielman et al. (2016). True site-wise and inferred dN/dS for their alignments can be found in their repository https://github.com/sjspielman/dnds_1rate_2rate

• filtered_sites contains information on all sites without any amino acid substitutions for each simulated alignment.

• processed_rates contains tables with all site-wise rates: true dN/dS, inferred dN/dS, and inferred Rate4Site.

• r4s_rates contains inferred site-wise Rate4site rates.

natural_prot contains results for the natural alignments from Spielman and Wilke (2013) and Meyer and Wilke (2015). The data we used can be found at https://github.com/sjspielman/mammalian_gpcr_selection and https://github.com/ausmeyer/hiv_structural_determinants, respectively.

• aln contains HIV-1 and GPCR protein sequences used in our analysis

  • aligned_seqs contains amino acid sequences we aligned.

  • back_translated_aln codon alignments that were translated back from amino acid alignments.

  • raw_aln contains raw FASTA files from the repositories mentioned.

  • reforematted_aln contains nucleotide alignments with sequence IDs reformatted. These were used as input for HyPhy.

• filtered_sites contains information on all sites without any amino acid substitutions for each alignment.

• inferred_dNdS contains site-wise inferred dN/dS.

• processed_rates contains tables with site-wise inferred dN/dS and inferred Rate4Site.

• r4s_rates contains inferred site-wise Rate4site rates.

• trees contains trees inferred from amino acid alignment for each protein. This directory also contains trees with reformatted sequence IDs to be used as input for HyPhy.

plots contains final figures used in the publication.

src contains all of the scripts used to analyze the data and plot the figures. The usage of each script is described in the section below.

Analysis

dN/dS model

The analysis in this section requires https://github.com/sjspielman/dnds_1rate_2rate in the same directory as the current repository.

1. Copy trees from https://github.com/sjspielman/dnds_1rate_2rate using the command line cp ../dnds_1rate_2rate/trees/n*_bl*.tre ./trees/.

2. Simulate alignments using ./src/write_run_sim_aln.sh. This script will write run_sim_aln.sh to simulate dN/dS alignments.

3. Translate simulated nucleotide alignments to amino acids using ./src/write_run_translate_aln.sh.

4. Infer site-wise dN/dS with HyPhy using the script ./src/dnds_inference/submit_run_inference.sh. This script was copied from https://github.com/sjspielman/dnds_1rate_2rate and modified for this analysis.

5. Infer site-wise Rate4Site scores using ./src/write_run_r4s_mech_codon.sh. This script will write run_r4s_mech_codon.sh which uses r4s_pipeline.sh to run Rate4Site on simulated alignments.

6. Concatenate all rates into a table with ./src/concatenate_mech_codon_rates.r.

MutSel model

The analysis in this section requires https://github.com/sjspielman/dnds_1rate_2rate in the same directory as the current repository.

1. Translate simulated nucleotide alignments from Spielman et al. (2016) to amino acids using ./src/write_run_translate_aln.sh.

2. Infer site-wise Rate4Site scores using ./src/write_run_r4s_mut_sel.sh. This script will write run_r4s_mut_sel.sh which uses r4s_pipeline.sh to run Rate4Site on simulated alignments.

3. Concatenate all rates into a table with ./src/concatenate_mut_sel_rates.r.

Natural proteins

1. Align amino acid sequences using ./src/write_run_align_natural_prot.sh.

2. Back translate amino acid alignments into codon alignments with ./src/run_translate_aln_aa_to_codon.sh. This script requires original nucleotide sequences.

3. Infer trees from the amino acid sequences with RAxML. The script ./src/write_run_raxml.sh will write run_raxml.sh which will run the inference.

4. Infer site-wise dN/dS with HyPhy using the script ./src/dnds_inference/submit_run_inference_nat_prot.sh. This script was copied from https://github.com/sjspielman/dnds_1rate_2rate and modified for this analysis.

5. Infer site-wise Rate4Site scores using ./src/write_run_r4s_natural_prot.sh. This script will write run_r4s_natural_prot.sh which will run Rate4Site on natural alignments.

6. Concatenate all rates into a table with ./src/concatenate_natural_prot_rates.r.