diff --git a/.gitignore b/.gitignore
index e2c51bd..28a38d1 100755
--- a/.gitignore
+++ b/.gitignore
@@ -86,7 +86,7 @@ results/summaries/*
!results/dms-viz
results/dms-viz/*
-!results/dms-viz/dms-viz.json
+!results/dms-viz/*.json
node_modules/
!homepage/.vitepress/
diff --git a/analysis_notebooks/configure_dms_viz.ipynb b/analysis_notebooks/configure_dms_viz.ipynb
index 3325a5d..2528cb9 100644
--- a/analysis_notebooks/configure_dms_viz.ipynb
+++ b/analysis_notebooks/configure_dms_viz.ipynb
@@ -266,7 +266,7 @@
"assert set(tooltip_cols).issubset(phenotypes.columns)\n",
"\n",
"filter_cols = [\"cell_entry\"]\n",
- "filter_limits = {\"cell_entry\": [phenotypes[\"cell_entry\"].min(), -3, 0]}"
+ "filter_limits = {\"cell_entry\": [float(phenotypes[\"cell_entry\"].min()), -3, 0]}"
]
},
{
@@ -364,7 +364,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
- "version": "3.11.9"
+ "version": "3.11.10"
}
},
"nbformat": 4,
diff --git a/analysis_notebooks/configure_dms_viz_mabs.ipynb b/analysis_notebooks/configure_dms_viz_mabs.ipynb
new file mode 100644
index 0000000..b1c44e7
--- /dev/null
+++ b/analysis_notebooks/configure_dms_viz_mabs.ipynb
@@ -0,0 +1,324 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "id": "6f4b8ac5-ff40-4449-a326-d63651e0d828",
+ "metadata": {},
+ "source": [
+ "# Configure structure based analysis for `dms-viz`"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "c371f68d-0e22-4188-8fca-11c78d6554d3",
+ "metadata": {},
+ "source": [
+ "Imports:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "52955e16-8de2-4385-a985-932a8a6a0beb",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "import gzip\n",
+ "import os\n",
+ "import requests\n",
+ "import subprocess\n",
+ "import textwrap\n",
+ "import warnings\n",
+ "\n",
+ "import Bio.PDB.PDBParser\n",
+ "import Bio.PDB.Polypeptide\n",
+ "\n",
+ "import matplotlib\n",
+ "\n",
+ "import pandas as pd\n",
+ "\n",
+ "import seaborn"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "27c25a7d-3439-460c-84c0-13599a4fbb80",
+ "metadata": {},
+ "source": [
+ "Define variables. This next cell is tagged `parameters` for `papermill` parameterization:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "0a26981c-1c55-4ed1-ab17-18ff6ca585b5",
+ "metadata": {
+ "editable": true,
+ "slideshow": {
+ "slide_type": ""
+ },
+ "tags": [
+ "parameters"
+ ]
+ },
+ "outputs": [],
+ "source": [
+ "pdb_id = None\n",
+ "phenotypes_csv = None\n",
+ "per_antibody_escape_csv = None\n",
+ "site_numbering_map = None\n",
+ "dms_viz_json = None\n",
+ "dms_viz_sitemap = None\n",
+ "dms_viz_phenotypes = None\n",
+ "pdb_file = None\n",
+ "dms_viz_subdir = None"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "36204609-74f7-488c-98dd-4da325d99f16",
+ "metadata": {},
+ "source": [
+ "Build the [sitemap](https://dms-viz.github.io/dms-viz-docs/preparing-data/data-requirements/#reference-site) used by `dms-viz`:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "6d5bfb2d-fed2-4213-86b9-e915f4ff9511",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "phenotypes = pd.read_csv(phenotypes_csv)\n",
+ "\n",
+ "sitemap = (\n",
+ " pd.read_csv(site_numbering_map)\n",
+ " .sort_values(\"sequential_site\")\n",
+ " .assign(\n",
+ " HA_chain=lambda x: x[\"HA1_HA2_H5_site\"].str.split().str[1].str[1: -1],\n",
+ " first_ha2_site=lambda x: x.query(\"HA_chain == 'HA2'\")[\"mature_H5_site\"].min(),\n",
+ " protein_site=lambda x: x[\"mature_H5_site\"].where(\n",
+ " x[\"HA_chain\"] == \"HA1\",\n",
+ " x[\"mature_H5_site\"] - x[\"first_ha2_site\"] + 1,\n",
+ " ),\n",
+ " chains=lambda x: x[\"HA_chain\"].map({\"HA1\": \"A\", \"HA2\": \"B\"}),\n",
+ " )\n",
+ " .merge(phenotypes[[\"site\", \"wildtype\"]].drop_duplicates().rename(columns={\"site\": \"reference_site\"}))\n",
+ " [[\"sequential_site\", \"reference_site\", \"protein_site\", \"wildtype\", \"HA_chain\", \"chains\"]]\n",
+ ")\n",
+ "\n",
+ "sitemap.to_csv(dms_viz_sitemap, index=False)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "fe8a3bb1-f557-43b2-9ce9-0418aaa63928",
+ "metadata": {},
+ "source": [
+ "Get the biological assembly (see https://pdb101.rcsb.org/learn/guide-to-understanding-pdb-data/biological-assemblies#Anchor-download) as the crystallographic unit doesn't correspond to that:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "4c2afd40-067f-462e-88d6-d3128dd57ff2",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "r = requests.get(f\"https://files.rcsb.org/download/{pdb_id}.pdb1.gz\")\n",
+ "assert r.status_code == 200\n",
+ "pdb_content = gzip.decompress(r.content).decode(\"utf-8\")\n",
+ "with open(pdb_file, \"w\") as f:\n",
+ " f.write(pdb_content)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "b6bbef2e-e52d-4735-9e8e-d5401ebb9f36",
+ "metadata": {},
+ "source": [
+ "Check the sites mismatched between the sitemap and the protein structure in terms of residue identity:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "7e1ec117-3166-471d-8485-602ab384b50f",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "with warnings.catch_warnings():\n",
+ " warnings.simplefilter(\"ignore\")\n",
+ " pdb_obj = Bio.PDB.PDBParser().get_structure(id=pdb_id, file=pdb_file)[0]\n",
+ "\n",
+ "records = []\n",
+ "for chain in [\"A\", \"B\"]:\n",
+ " for res in pdb_obj[chain].get_residues():\n",
+ " if not res.id[0].isspace():\n",
+ " continue\n",
+ " aa = Bio.PDB.Polypeptide.protein_letters_3to1[res.resname]\n",
+ " r = res.id[1]\n",
+ " records.append((chain, r, aa))\n",
+ "pdb_df = pd.DataFrame(records, columns=[\"chains\", \"protein_site\", \"pdb_aa\"])\n",
+ "\n",
+ "mismatched_sites = sitemap.merge(pdb_df, how=\"left\")\n",
+ "\n",
+ "print(\n",
+ " f\"Of {len(sitemap)} sites, {len(mismatched_sites.query('wildtype == pdb_aa'))} match, \"\n",
+ " f\"{len(mismatched_sites.query('pdb_aa.isnull()'))} are missing from PDB, and \"\n",
+ " f\"{len(mismatched_sites.query('pdb_aa.notnull()').query('wildtype != pdb_aa'))} differ.\"\n",
+ ")\n",
+ "\n",
+ "print(\"Sites that differ:\")\n",
+ "display(mismatched_sites.query(\"pdb_aa.notnull() and (wildtype != pdb_aa)\").reset_index(drop=True))"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "9f5dd45f-1627-425d-8cb2-63f5f103dcea",
+ "metadata": {},
+ "source": [
+ "Write the phenotypes after adding the antibodyd escape to a CSV file:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "562538b8-d250-40a9-ab01-6468bcccafa9",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "per_antibody_escape = (\n",
+ " pd.read_csv(per_antibody_escape_csv)\n",
+ " .drop(columns=\"antibody_set\")\n",
+ " .merge(\n",
+ " phenotypes[\n",
+ " [\n",
+ " \"site\",\n",
+ " \"mutant\",\n",
+ " \"entry in 293T cells\",\n",
+ " \"sequential_site\",\n",
+ " \"mature_H5_site\",\n",
+ " \"HA1_HA2_H5_site\",\n",
+ " ]\n",
+ " ],\n",
+ " on=[\"site\", \"mutant\"],\n",
+ " validate=\"many_to_one\",\n",
+ " how=\"left\",\n",
+ " )\n",
+ " .rename(columns={\"entry in 293T cells\": \"cell_entry\"})\n",
+ " .assign(\n",
+ " mutation=lambda x: x[\"wildtype\"] + x[\"site\"].astype(str) + x[\"mutant\"],\n",
+ " )\n",
+ ")\n",
+ "\n",
+ "antibodies = list(per_antibody_escape[\"antibody\"].unique())\n",
+ "\n",
+ "print(f\"Read escape for {antibodies=}\")\n",
+ "\n",
+ "print(f\"Writing the phenotypes to {dms_viz_phenotypes}\")\n",
+ "per_antibody_escape.to_csv(dms_viz_phenotypes, index=False, float_format=\"%.4g\")\n",
+ "\n",
+ "print(f\"{per_antibody_escape.columns=}\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "e7ff0c20-4fc0-497d-b114-1f50ea042fbb",
+ "metadata": {},
+ "source": [
+ "Get enough colors:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "ce9e2f01-bb4a-48e9-a99c-3d5d73da0d72",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "def get_hex_color_palette(num_colors):\n",
+ " colors = seaborn.color_palette(\"hls\", num_colors)\n",
+ " hex_colors = [matplotlib.colors.to_hex(color) for color in colors]\n",
+ " return hex_colors\n",
+ "\n",
+ "nconditions = len(antibodies)\n",
+ "if nconditions > 4:\n",
+ " colors = \",\".join(get_hex_color_palette(nconditions))\n",
+ "else:\n",
+ " colors = \"#0072B2,#CC79A7,#4C3549,#009E73\"\n",
+ "\n",
+ "print(f\"Using {colors=}\")\n",
+ "seaborn.palplot(seaborn.color_palette(colors.split(\",\")))"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "00e1953c-bd3f-40fc-83a5-84233be74526",
+ "metadata": {},
+ "source": [
+ "Run [configure-dms-viz](https://dms-viz.github.io/dms-viz-docs/preparing-data/command-line-api/).\n",
+ "First, set up some options:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "22448e4d-a621-49cb-aabc-81baaf821f88",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "tooltip_cols = [\n",
+ " \"mutation\",\n",
+ " \"sequential_site\",\n",
+ " \"mature_H5_site\",\n",
+ " \"cell_entry\",\n",
+ "]\n",
+ "\n",
+ "cmds = [\n",
+ " \"configure-dms-viz\", \"format\",\n",
+ " \"--name\", \"VRC antibodies\",\n",
+ " \"--input\", dms_viz_phenotypes,\n",
+ " \"--output\", dms_viz_json,\n",
+ " \"--structure\", pdb_file,\n",
+ " \"--metric\", \"escape\",\n",
+ " \"--condition\", \"antibody\",\n",
+ " \"--sitemap\", dms_viz_sitemap,\n",
+ " \"--colors\", colors,\n",
+ " \"--alphabet\", \"RKHDEQNSTYWFAILMVGPC\",\n",
+ " \"--summary-stat\", \"sum\",\n",
+ " \"--floor\", \"True\",\n",
+ " \"--tooltip-cols\", str({c: c.replace(\"_\", \" \") for c in tooltip_cols}),\n",
+ " \"--filter-cols\", \"{'cell_entry': 'cell entry'}\",\n",
+ " \"--filter-limits\", f\"{{'cell_entry': [{float(per_antibody_escape['cell_entry'].min())}, -3, 0]}}\",\n",
+ " \"--title\", \"H5 HA escape from VRC antibodies as measured by pseudovirus deep mutational scanning\",\n",
+ " \"--description\", \"H5 HA escape from VRC antibodies as measured by pseudovirus deep mutational scanning\",\n",
+ "]\n",
+ "\n",
+ "print(f\"Running the following commands:\\n{cmds}\")\n",
+ "subprocess.run(cmds, check=True)"
+ ]
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "Python 3 (ipykernel)",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.11.10"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/custom_rules.smk b/custom_rules.smk
index 09d25d3..18745f4 100755
--- a/custom_rules.smk
+++ b/custom_rules.smk
@@ -135,6 +135,42 @@ rule configure_dms_viz:
&> {log}
"""
+
+rule configure_dms_viz_vrc_mabs:
+ """Configure a JSON file for `dms-viz` for the VRC antibodies"""
+ input:
+ phenotypes_csv="results/summaries/phenotypes.csv",
+ per_antibody_escape_csv="results/summaries/VRC_antibody_escape_per_antibody_escape.csv",
+ site_numbering_map=config["site_numbering_map"],
+ nb="analysis_notebooks/configure_dms_viz_mabs.ipynb",
+ output:
+ dms_viz_json="results/dms-viz/dms-viz_vrc_mabs.json",
+ dms_viz_sitemap="results/dms-viz/sitemap_vrc_mabs.csv",
+ dms_viz_phenotypes="results/dms-viz/phenotypes_vrc_mabs.csv",
+ pdb_file="results/dms-viz/pdb_file_vrc_mabs.pdb",
+ nb="results/notebooks/configure_dms_viz_vrc_mabs.ipynb",
+ params:
+ dms_viz_subdir=lambda _, output: os.path.dirname(output.dms_viz_json),
+ pdb_id="4kwm",
+ log:
+ "results/logs/configure_dms_viz_vrc_mabs.txt",
+ conda:
+ "envs/dms-viz.yml"
+ shell:
+ """
+ papermill {input.nb} {output.nb} \
+ -p phenotypes_csv {input.phenotypes_csv} \
+ -p per_antibody_escape_csv {input.per_antibody_escape_csv} \
+ -p site_numbering_map {input.site_numbering_map} \
+ -p dms_viz_json {output.dms_viz_json} \
+ -p dms_viz_sitemap {output.dms_viz_sitemap} \
+ -p dms_viz_phenotypes {output.dms_viz_phenotypes} \
+ -p pdb_file {output.pdb_file} \
+ -p dms_viz_subdir {params.dms_viz_subdir} \
+ -p pdb_id {params.pdb_id} \
+ &> {log}
+ """
+
# Files (Jupyter notebooks, HTML plots, or CSVs) that you want included in
# the HTML docs should be added to the nested dict `docs`:
docs["Additional files"] = {
@@ -150,6 +186,8 @@ docs["Additional files"] = {
"Notebook plotting functional effect distribution":
rules.functional_effect_distribution.output.nb,
},
+ "Visualizations of DMS data on protein structure (dms-viz JSONs)": {
+ "Dadonaite (2024) phenotypes JSON": rules.configure_dms_viz.output.dms_viz_json,
+ "VRC antibodies JSON": rules.configure_dms_viz_vrc_mabs.output.dms_viz_json,
+ },
}
-
-other_target_files.append(rules.configure_dms_viz.output.dms_viz_json)
diff --git a/docs/index.html b/docs/index.html
index 8bf6ab7..671a1c5 100644
--- a/docs/index.html
+++ b/docs/index.html
@@ -732,4 +732,9 @@ Cell entry effect comparison
Functional effect distribution
+Visualizations of DMS data on protein structure (dms-viz JSONs)
+
\ No newline at end of file
diff --git a/envs/dms-viz.yml b/envs/dms-viz.yml
index 537efff..6033045 100755
--- a/envs/dms-viz.yml
+++ b/envs/dms-viz.yml
@@ -7,10 +7,11 @@ channels:
dependencies:
- biopython=1.83
- jupyterlab
- - pandas=2.1 # required because of https://github.com/matsengrp/multidms/issues/128
- - papermill=2.4
+ - pandas=2.2
+ - papermill=2.6
- pip
- pyarrow
- python=3.11
+ - seaborn=0.13
- pip:
- configure-dms-viz==1.6.0
diff --git a/homepage/public/appendix.html b/homepage/public/appendix.html
index 8bf6ab7..671a1c5 100644
--- a/homepage/public/appendix.html
+++ b/homepage/public/appendix.html
@@ -732,4 +732,9 @@ Cell entry effect comparison
Functional effect distribution
+Visualizations of DMS data on protein structure (dms-viz JSONs)
+