Dimer Analysis – Code and Data Overview

This repository contains code, input data, and figures used in the manuscript
“Differential conservation analysis identifies residues defining constitutive internalization in beta‑adrenergic receptors” .

There are three main components:

• Ortholog multiple sequence alignment (MSA) for calculating residue conservation within β-adrenergic receptor subtypes.
• Structural visualization of the TM1/TM7/H8 dimer interface (Figure 1B), which is independent of RRCS.
• Quantitative analysis of dimerization interface contacts from MD simulations using the residue–residue contact score (RRCS) algorithm of Zhou and co‑workers, implemented following Jones et al., eLife 2020 https://doi.org/10.7554/eLife.54895.

Ortholog Multiple Sequence Alignment (MSA)

• BAR_orthologs_MSA.fasta
  • Multiple sequence alignment of β-adrenergic receptor orthologs used to calculate residue conservation within orthologs of each β-adrenergic receptor subtype.
  • The file is organized in a structured format where:
    • Human sequences (ADRB1_HUMAN, ADRB2_HUMAN, ADRB3_HUMAN) appear first for each subtype
    • Ortholog sequences from other species immediately follow each human sequence
    • This pattern repeats for all three β-adrenergic receptor subtypes (β₁AR, β₂AR, β₃AR)

Example structure of the fasta file:

>sp|P08588|ADRB1_HUMAN|9606/1-477          # Human β₁AR
[alignment sequence...]
>tr|G3SE09|G3SE09_GORGO|9595/1-291         # Gorilla β₁AR ortholog
[alignment sequence...]
>tr|...|..._PANTR|...                       # Chimpanzee β₁AR ortholog
[alignment sequence...]
[... additional β₁AR orthologs ...]

>sp|P07550|ADRB2_HUMAN|9606/1-413          # Human β₂AR
[alignment sequence...]
[... β₂AR orthologs ...]

>sp|P13945|ADRB3_HUMAN|9606/1-408          # Human β₃AR
[alignment sequence...]
[... β₃AR orthologs ...]

This organization enables subtype-specific conservation analysis by grouping each human receptor with its respective orthologs.

Structural visualization for Figure 1B (independent of RRCS)

• structural_visualization_adrb1_turkey.pse

  • PyMOL session used to generate structural views of the BAR dimer based on the ligand‑free turkey β1‑adrenergic receptor template.
  • Encodes the representation, coloring, selection of TM1/TM7/H8, and orientation used in the manuscript.

• dimer_figure1B.png

  • Exported PyMOL image from the above session.
  • Used as Figure 1B in the manuscript to illustrate the TM1/TM7/H8 dimer interface and the differentially conserved residues.

RRCS‑based dimer interface analysis (Figure 3)

Core RRCS implementation

• RRCS.py
  • Stand‑alone Python implementation of the RRCS algorithm.
  • Takes a single PDB file as input and writes a corresponding *.cscore text file.
  • Each line of the output lists a residue pair and its RRCS value:
    CHAIN:RESNUM_RESNAME CHAIN:RESNUM_RESNAME CONTACT_SCORE.

Batch scoring of MD snapshots

• rrcs_complier.py
  • Convenience script to run RRCS.py on many MD snapshots.
  • Assumes two main snapshot directories on disk:
    • Active-Inactive
    • Inactive-Inactive (mirrored here under Inactive-Inactive/WT, Inactive-Inactive/V34A, Inactive-Inactive/S41A, Inactive-Inactive/V34A_S41A, Inactive-Inactive/F49A).
  • Recursively finds all *.pdb snapshots in these folders and executes
    python RRCS.py <snapshot.pdb>
    to generate matching *.cscore files for every frame and mutant.

MD snapshot and RRCS data layout

• Inactive-Inactive/
  • Contains subfolders for each mutant (WT, V34A, S41A, V34A_S41A, F49A).
  • Each mutant folder holds multiple trajectory replicas, e.g. WT_II_1_PDB, WT_II_2_PDB, …
  • Within each replica folder:
    • *.pdb — individual MD snapshots of the dimer.
    • *.pdb.cscore — RRCS output from RRCS.py for the corresponding snapshot.
  • These .cscore files are the raw input for the downstream total interface score calculations.

Aggregating interface contact scores

• total_rrcs_dimer_boxplot.py
  • Walks the project directory tree, reads every *.cscore file, and extracts A–B dimer interface contact scores.
  • For each snapshot, it sums all RRCS values between chain A and chain B residues to obtain a single total interface contact score.
  • Aggregates these totals by mutation (e.g. WT_II, V34A_II, S41A_II, V34A_S41A_II, F49A_II).
  • Writes a long‑format CSV:
    • boxplot_data_II.csv – columns: Mutation, Contact Score; each row is one MD frame’s A–B interface total.

Statistical analysis and visualization

• interface_contact_score_boxplot.R
  • R script that reads boxplot_data_II.csv and prepares Figure 3 of the manuscript.
  • Steps:
    • Recodes the Mutation factor to human‑readable labels (WT, V34A, S41A, V34A-S41A, F49A).
    • Performs one‑way ANOVA on Contact.Score ~ Mutation and, if significant, Tukey HSD post‑hoc tests.
    • Produces a ggplot2 box‑and‑jitter plot of interface contact scores by mutation.
  • Saves the final vector figure as:
    • MD_boxplot.svg – boxplot of total dimer interface contact scores (used as Figure 3 in the manuscript).

Typical workflow

1. Generate RRCS scores for MD snapshots
   • Use rrcs_complier.py (or manually call RRCS.py) on all PDB snapshots in the MD folders to produce *.cscore files.
2. Aggregate A–B interface scores
   • Run total_rrcs_dimer_boxplot.py to create boxplot_data_II.csv from the .cscore outputs.
3. Analyze and plot
   • Run interface_contact_score_boxplot.R to perform ANOVA/Tukey tests and generate MD_boxplot.svg (Figure 3).
4. Structural figure generation
   • Open structural_visualization_adrb1_turkey.pse in PyMOL and export dimer_figure1B.png (Figure 1B).