OSCR

Candidalysin Inhibits <i>Porphyromonas gingivalis</i> Lipoprotein-Induced IL-1β Production in BV-2 Microglia via Hydrophobic Microbial Interactions.

Code ↔ Paper

2 matches between paragraphs of the paper and lines of its authors' code, computed by the harvester (lexical-v1). Click a colored paragraph or line to see its counterpart.

The 2 matches
  1. [1] § 4. Materials and Methods › 4.9. AlphaFold Predictions ↔ beta/ESMFold_advanced.ipynb, lines 99–245 · score 0.57 · amino acid sequence, best model, recycles, predicted
  2. [2] § 4. Materials and Methods › 4.9. AlphaFold Predictions ↔ colabfold/citations.py, lines 1–60 · score 0.55 · AlphaFold2, ColabFold, server, acid, predicted, protein

Paper

Loaded from Europe PMC by your browser, not stored by OSCR: doi.org · Europe PMC

The paper is loaded when this pane is shown.

The authors' code

Jupyter notebook · 496 lines · 17 KB · MIT · 1 match

  1. # %% [markdown]
  2. # <a href="https://colab.research.google.com/github/sokrypton/ColabFold/blob/main/beta/ESMFold_advanced.ipynb" target="_parent"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/></a>
  4. # %% [markdown]
  5. # #**ESMFold_advanced**
  6. # for more details see: [Github](https://github.com/facebookresearch/esm/tree/main/esm), [Preprint](https://www.biorxiv.org/content/10.1101/2022.07.20.500902v1)
  7. #
  8. # #### **Tips and Instructions**
  9. # - click the little ▶ play icon to the left of each cell below.
  10. #
  11. # #### **Colab Limitations**
  12. # - On Tesla T4 (typical free colab GPU), max total length ~ 900
  13. #
  15. # %%
  16. %%time
  17. #@title install
  18. #@markdown install ESMFold, OpenFold and download Params (~2min 30s)
  20. import os, time
  21. import torch
  22. if not os.path.isfile("esmfold.model"):
  23. # download esmfold params
  24. os.system("apt-get install aria2 -qq")
  25. os.system("aria2c -q -x 16 https://colabfold.steineggerlab.workers.dev/esm/esmfold.model &")
  27. # install libs
  28. print("installing libs...")
  29. os.system("pip install -q omegaconf pytorch_lightning biopython ml_collections einops py3Dmol modelcif")
  30. os.system("pip install -q git+https://github.com/NVIDIA/dllogger.git")
  32. print("installing openfold...")
  33. # install openfold
  34. os.system(f"pip install -q git+https://github.com/sokrypton/openfold.git")
  36. print("installing esmfold...")
  37. # install esmfold
  38. os.system(f"pip install -q git+https://github.com/sokrypton/esm.git@beta")
  40. # wait for Params to finish downloading...
  41. if not os.path.isfile("esmfold.model"):
  42. # backup source!
  43. os.system("aria2c -q -x 16 https://files.ipd.uw.edu/pub/esmfold/esmfold.model")
  44. else:
  45. while os.path.isfile("esmfold.model.aria2"):
  46. time.sleep(5)
  48. if "model" not in dir():
  49. model_path = "esmfold.model"
  50. model = torch.load(model_path, weights_only=False)
  51. model.cuda().requires_grad_(False)
  53. import os
  54. import re
  55. import hashlib
  56. import numpy as np
  57. import torch
  58. from string import ascii_uppercase, ascii_lowercase
  59. from jax.tree_util import tree_map
  60. from scipy.special import softmax
  61. from typing import Optional
  63. def get_hash(x: str) -> str:
  64. """Generates a SHA1 hash for a given string."""
  65. return hashlib.sha1(x.encode()).hexdigest()
  67. def parse_output(output: dict) -> dict:
  68. """
  69. Parses the raw model output dictionary to extract key metrics and structures.
  71. Args:
  72. output: The raw dictionary from model.infer().
  74. Returns:
  75. A dictionary containing pae, plddt, contacts, and xyz coordinates.
  76. """
  77. pae = (output["aligned_confidence_probs"][0] * np.arange(64)).mean(-1) * 31
  78. plddt = output["plddt"][0, :, 1]
  80. bins = np.append(0, np.linspace(2.3125, 21.6875, 63))
  81. sm_contacts = softmax(output["distogram_logits"], -1)[0]
  82. sm_contacts = sm_contacts[..., bins < 8].sum(-1)
  83. xyz = output["positions"][-1, 0, :, 1]
  84. mask = output["atom37_atom_exists"][0, :, 1] == 1
  86. o = {"pae": pae[mask, :][:, mask],
  87. "plddt": plddt[mask],
  88. "sm_contacts": sm_contacts[mask, :][:, mask],
  89. "xyz": xyz[mask]}
  91. if "contacts" in output["lm_output"]:
  92. lm_contacts = output["lm_output"]["contacts"].astype(float)[0]
  93. o["lm_contacts"] = lm_contacts[mask, :][:, mask]
  95. return o
  97. # --- Main Prediction Function ---
  99. def predict_esmfold(
  100. model: torch.nn.Module,
  101. sequence: str,
  102. jobname: str = "test",
  103. num_recycles: int = 3,
  104. get_LM_contacts: bool = False,
  105. copies: int = 1,
  106. chain_linker: int = 25,
  107. samples: Optional[int] = None,
  108. masking_rate: float = 0.15,
  109. stochastic_mode: str = "LM",
  110. save_dir: str = ".",
  111. verbose=False
  112. ) -> dict:
  113. """
  114. Runs ESMFold prediction for a given sequence and parameters.
  116. Args:
  117. model: The loaded ESMFold torch model.
  118. sequence: The amino acid sequence (e.g., "GWSTELEKH...").
  119. jobname: A name for the prediction job.
  120. num_recycles: Number of recycling iterations (0, 1, 2, 3, 6, 12).
  121. get_LM_contacts: Whether to return language model contacts.
  122. copies: Number of copies for homo-oligomeric predictions.
  123. chain_linker: Length of the 'X' linker for multimers.
  124. samples: Number of stochastic samples (None, 1, 4, 8, 16, 32, 64).
  125. masking_rate: Masking rate for stochastic LM sampling.
  126. stochastic_mode: Type of stochastic sampling ("LM", "LM_SM", "SM").
  127. save_dir: Directory to save PDB files and results.
  129. Returns:
  130. A dictionary containing:
  131. - "best_output": Raw output dictionary of the best model.
  132. - "best_pdb_str": PDB string of the best model (highest ptm).
  133. - "best_ptm": The ptm score of the best model.
  134. - "trajectory": A list of parsed outputs for each sample.
  135. - "prediction_dir": The specific directory where results were saved.
  136. """
  138. # --- Input Processing ---
  139. jobname = re.sub(r'\W+', '', jobname)[:50]
  141. # Clean sequence
  142. sequence = re.sub("[^A-Z:]", "", sequence.replace("/", ":").upper())
  143. sequence = re.sub(":+", ":", sequence)
  144. sequence = re.sub("^[:]+", "", sequence)
  145. sequence = re.sub("[:]+$", "", sequence)
  147. if copies <= 0:
  148. copies = 1
  149. sequence = ":".join([sequence] * copies)
  151. seqs = sequence.split(":")
  152. lengths = [len(s) for s in seqs]
  153. length = sum(lengths)
  154. if verbose:
  155. print(f"Running prediction for job '{jobname}' with length {length}")
  157. ID = jobname + "_" + get_hash(sequence)[:5]
  158. prediction_dir = os.path.join(save_dir, ID)
  159. os.makedirs(prediction_dir, exist_ok=True)
  161. # --- Model Configuration ---
  162. # Optimized for Tesla T4
  163. if length > 700:
  164. model.trunk.set_chunk_size(64)
  165. else:
  166. model.trunk.set_chunk_size(128)
  168. if not next(model.parameters()).is_cuda:
  169. print("Warning: Model is not on CUDA. Moving to CUDA.")
  170. model.cuda()
  171. model.requires_grad_(False)
  173. # --- Prediction Loop ---
  174. best_pdb_str = None
  175. best_ptm = 0
  176. best_output = None
  177. traj = []
  179. num_samples = 1 if samples is None else samples
  181. for seed in range(num_samples):
  182. torch.cuda.empty_cache()
  184. current_seed_str = "default"
  185. mask_rate = 0.0
  186. model.train(False) # Set to eval mode by default
  188. if samples is not None:
  189. torch.manual_seed(seed)
  190. current_seed_str = f"seed{seed}"
  191. mask_rate = masking_rate if "LM" in stochastic_mode else 0.0
  192. model.train("SM" in stochastic_mode) # Set to train mode for dropout
  194. if verbose:
  195. print(f"Running sample {seed + 1}/{num_samples} (seed: {current_seed_str})...")
  197. output = model.infer(
  198. sequence,
  199. num_recycles=num_recycles,
  200. chain_linker="X" * chain_linker,
  201. residue_index_offset=512,
  202. mask_rate=mask_rate,
  203. return_contacts=get_LM_contacts
  204. )
  206. pdb_str = model.output_to_pdb(output)[0]
  207. output = tree_map(lambda x: x.cpu().numpy(), output)
  208. ptm = output["ptm"][0]
  209. plddt = output["plddt"][0, :, 1].mean()
  210. traj.append(parse_output(output))
  212. if verbose:
  213. print(f'Sample {seed} ptm: {ptm:.3f} plddt: {plddt:.1f}')
  215. if ptm > best_ptm:
  216. best_pdb_str = pdb_str
  217. best_ptm = ptm
  218. best_output = output
  220. # --- Save PDB File ---
  221. if samples is None:
  222. pdb_filename = f"ptm{ptm:.3f}_r{num_recycles}_{current_seed_str}.pdb"
  223. else:
  224. pdb_filename = (
  225. f"ptm{ptm:.3f}_r{num_recycles}_{current_seed_str}_"
  226. f"{stochastic_mode}_m{masking_rate:.2f}.pdb"
  227. )
  229. pdb_path = os.path.join(prediction_dir, pdb_filename)
  230. with open(pdb_path, "w") as f:
  231. f.write(pdb_str)
  232. if verbose:
  233. print(f"Saved PDB to {pdb_path}")
  235. if verbose:
  236. print(f"Prediction complete. Best ptm: {best_ptm:.3f}")
  238. return {
  239. "output": best_output,
  240. "pdb_str": best_pdb_str,
  241. "ptm": best_ptm,
  242. "traj": traj,
  243. "dir": prediction_dir,
  244. "lengths":lengths
  245. }
  248. # %%
  249. #@title ##run **ESMFold**
  250. jobname = "test" #@param {type:"string"}
  251. sequence = "GWSTELEKHREELKEFLKKEGITNVEIRIDNGRLEVRVEGGTERLKRFLEELRQKLEKKGYTVDIKIE" #@param {type:"string"}
  253. #@markdown ---
  254. #@markdown ###**Advanced Options**
  255. num_recycles = 3 #@param ["0", "1", "2", "3", "6", "12"] {type:"raw"}
  256. get_LM_contacts = False #@param {type:"boolean"}
  258. #@markdown **multimer options (experimental)**
  259. #@markdown - use "/" to specify chainbreaks, (eg. sequence="AAA/AAA")
  260. #@markdown - for homo-oligomeric predictions, set copies > 1
  262. copies = 1 #@param {type:"integer"}
  263. chain_linker = 25 #@param {type:"number"}
  265. #@markdown **sampling options (experimental)**
  266. #@markdown - Samples are generated via random masking (defined by `masking_rate`)
  267. #@markdown of input sequence (stochastic_mode="LM") and/or via dropout within structure module (stochastic_mode="SM").
  268. samples = None #@param ["None", "1", "4", "8", "16", "32", "64"] {type:"raw"}
  269. masking_rate = 0.15 #@param {type:"number"}
  270. stochastic_mode = "LM" #@param ["LM", "LM_SM", "SM"]
  272. o = predict_esmfold(
  273. model=model,
  274. sequence=sequence,
  275. jobname=jobname,
  276. num_recycles=num_recycles,
  277. get_LM_contacts=get_LM_contacts,
  278. copies=copies,
  279. chain_linker=chain_linker,
  280. samples=samples,
  281. masking_rate=masking_rate,
  282. stochastic_mode=stochastic_mode)
  284. # %%
  285. #@title display (optional) {run: "auto"}
  286. #@markdown Note: If samples selected, the model with max pTM is displayed.
  287. import py3Dmol
  290. pymol_color_list = ["#33ff33","#00ffff","#ff33cc","#ffff00","#ff9999","#e5e5e5","#7f7fff","#ff7f00",
  291. "#7fff7f","#199999","#ff007f","#ffdd5e","#8c3f99","#b2b2b2","#007fff","#c4b200",
  292. "#8cb266","#00bfbf","#b27f7f","#fcd1a5","#ff7f7f","#ffbfdd","#7fffff","#ffff7f",
  293. "#00ff7f","#337fcc","#d8337f","#bfff3f","#ff7fff","#d8d8ff","#3fffbf","#b78c4c",
  294. "#339933","#66b2b2","#ba8c84","#84bf00","#b24c66","#7f7f7f","#3f3fa5","#a5512b"]
  296. def show_pdb(pdb_str, show_sidechains=False, show_mainchains=False,
  297. color="pLDDT", chains=None, vmin=50, vmax=90,
  298. size=(800,480), hbondCutoff=4.0,
  299. Ls=None,
  300. animate=False):
  302. if chains is None:
  303. chains = 1 if Ls is None else len(Ls)
  304. view = py3Dmol.view(js='https://3dmol.org/build/3Dmol.js', width=size[0], height=size[1])
  305. if animate:
  306. view.addModelsAsFrames(pdb_str,'pdb',{'hbondCutoff':hbondCutoff})
  307. else:
  308. view.addModel(pdb_str,'pdb',{'hbondCutoff':hbondCutoff})
  309. if color == "pLDDT":
  310. view.setStyle({'cartoon': {'colorscheme': {'prop':'b','gradient': 'roygb','min':vmin,'max':vmax}}})
  311. elif color == "rainbow":
  312. view.setStyle({'cartoon': {'color':'spectrum'}})
  313. elif color == "chain":
  314. for n,chain,color in zip(range(chains),alphabet_list,pymol_color_list):
  315. view.setStyle({'chain':chain},{'cartoon': {'color':color}})
  316. if show_sidechains:
  317. BB = ['C','O','N']
  318. view.addStyle({'and':[{'resn':["GLY","PRO"],'invert':True},{'atom':BB,'invert':True}]},
  319. {'stick':{'colorscheme':f"WhiteCarbon",'radius':0.3}})
  320. view.addStyle({'and':[{'resn':"GLY"},{'atom':'CA'}]},
  321. {'sphere':{'colorscheme':f"WhiteCarbon",'radius':0.3}})
  322. view.addStyle({'and':[{'resn':"PRO"},{'atom':['C','O'],'invert':True}]},
  323. {'stick':{'colorscheme':f"WhiteCarbon",'radius':0.3}})
  324. if show_mainchains:
  325. BB = ['C','O','N','CA']
  326. view.addStyle({'atom':BB},{'stick':{'colorscheme':f"WhiteCarbon",'radius':0.3}})
  327. view.zoomTo()
  328. if animate: view.animate()
  329. return view
  331. color = "confidence" #@param ["confidence", "rainbow", "chain"]
  332. if color == "confidence": color = "pLDDT"
  333. show_sidechains = False #@param {type:"boolean"}
  334. show_mainchains = False #@param {type:"boolean"}
  335. show_pdb(o["pdb_str"], color=color,
  336. show_sidechains=show_sidechains,
  337. show_mainchains=show_mainchains,
  338. Ls=o["lengths"]).show()
  340. # %%
  341. #@title plot confidence (optional)
  342. #@markdown Note: If samples selected, the model with max pTM is displayed.
  343. dpi = 100 #@param {type:"integer"}
  345. import matplotlib.pyplot as plt
  346. def plot_ticks(Ls):
  347. Ln = sum(Ls)
  348. L_prev = 0
  349. for L_i in Ls[:-1]:
  350. L = L_prev + L_i
  351. L_prev += L_i
  352. plt.plot([0,Ln],[L,L],color="black")
  353. plt.plot([L,L],[0,Ln],color="black")
  354. ticks = np.cumsum([0]+Ls)
  355. ticks = (ticks[1:] + ticks[:-1])/2
  356. plt.yticks(ticks,alphabet_list[:len(ticks)])
  358. def plot_confidence(output, Ls=None, dpi=100):
  359. O = parse_output(output)
  360. if "lm_contacts" in O:
  361. plt.figure(figsize=(20,4), dpi=dpi)
  362. plt.subplot(1,4,1)
  363. else:
  364. plt.figure(figsize=(15,4), dpi=dpi)
  365. plt.subplot(1,3,1)
  367. plt.title('Predicted lDDT')
  368. plt.plot(O["plddt"])
  369. if Ls is not None:
  370. L_prev = 0
  371. for L_i in Ls[:-1]:
  372. L = L_prev + L_i
  373. L_prev += L_i
  374. plt.plot([L,L],[0,100],color="black")
  375. plt.xlim(0,O["plddt"].shape[0])
  376. plt.ylim(0,100)
  377. plt.ylabel('plDDT')
  378. plt.xlabel('position')
  379. plt.subplot(1,4 if "lm_contacts" in O else 3,2)
  381. plt.title('Predicted Aligned Error')
  382. Ln = O["pae"].shape[0]
  383. plt.imshow(O["pae"],cmap="bwr",vmin=0,vmax=30,extent=(0, Ln, Ln, 0))
  384. if Ls is not None and len(Ls) > 1: plot_ticks(Ls)
  385. plt.colorbar()
  386. plt.xlabel('Scored residue')
  387. plt.ylabel('Aligned residue')
  389. if "lm_contacts" in O:
  390. plt.subplot(1,4,3)
  391. plt.title("contacts from LM")
  392. plt.imshow(O["lm_contacts"],cmap="Greys",vmin=0,vmax=1,extent=(0, Ln, Ln, 0))
  393. if Ls is not None and len(Ls) > 1: plot_ticks(Ls)
  394. plt.subplot(1,4,4)
  395. else:
  396. plt.subplot(1,3,3)
  397. plt.title("contacts from Structure Module")
  398. plt.imshow(O["sm_contacts"],cmap="Greys",vmin=0,vmax=1,extent=(0, Ln, Ln, 0))
  399. if Ls is not None and len(Ls) > 1: plot_ticks(Ls)
  400. return plt
  402. plot_confidence(o["output"], Ls=o["lengths"], dpi=dpi)
  403. plt.show()
  405. # %%
  406. #@title download predictions
  407. from google.colab import files
  408. os.system(f"zip {o['dir']}.zip {o['dir']}/*")
  409. files.download(f"{o['dir']}.zip")
  411. # %%
  412. #@title animate outputs (optional)
  413. #@markdown if more than one sample generated
  414. import os
  415. if not os.path.isdir("colabfold"):
  416. os.system("git clone https://github.com/sokrypton/ColabFold.git")
  417. import ColabFold.beta.colabfold as cf
  419. dpi = 100#@param {type:"integer"}
  420. color_by_plddt = True #@param {type:"boolean"}
  421. use_pca = True
  422. cycle = True
  424. import matplotlib
  425. from matplotlib import animation
  426. from IPython.display import HTML
  427. from sklearn.decomposition import PCA
  429. def mk_animation(xyz, labels, plddt, ref=0, Ls=None, line_w=2.0, dpi=100,color_by_plddt=False):
  431. def ca_align_to_last(positions, ref):
  432. def align(P, Q):
  433. if Ls is None or len(Ls) == 1:
  434. P_,Q_ = P,Q
  435. else:
  436. # align relative to first chain
  437. P_,Q_ = P[:Ls[0]],Q[:Ls[0]]
  438. p = P_ - P_.mean(0,keepdims=True)
  439. q = Q_ - Q_.mean(0,keepdims=True)
  440. return ((P - P_.mean(0,keepdims=True)) @ cf.kabsch(p,q)) + Q_.mean(0,keepdims=True)
  442. pos = positions[ref] - positions[ref].mean(0,keepdims=True)
  443. best_2D_view = pos @ cf.kabsch(pos,pos,return_v=True)
  445. new_positions = []
  446. for i in range(len(positions)):
  447. new_positions.append(align(positions[i],best_2D_view))
  448. return np.asarray(new_positions)
  450. # align to reference
  451. pos = ca_align_to_last(xyz, ref)
  453. fig, (ax1) = plt.subplots(1)
  454. fig.set_figwidth(5)
  455. fig.set_figheight(5)
  456. fig.set_dpi(dpi)
  458. xy_min = pos[...,:2].min() - 1
  459. xy_max = pos[...,:2].max() + 1
  461. for ax in [ax1]:
  462. ax.set_xlim(xy_min, xy_max)
  463. ax.set_ylim(xy_min, xy_max)
  464. ax.axis(False)
  466. ims=[]
  467. for l,pos_,plddt_ in zip(labels,pos,plddt):
  468. if color_by_plddt:
  469. img = cf.plot_pseudo_3D(pos_, c=plddt_, cmin=50, cmax=90, line_w=line_w, ax=ax1)
  470. elif Ls is None or len(Ls) == 1:
  471. img = cf.plot_pseudo_3D(pos_, ax=ax1, line_w=line_w)
  472. else:
  473. c = np.concatenate([[n]*L for n,L in enumerate(Ls)])
  474. img = cf.plot_pseudo_3D(pos_, c=c, cmap=cf.pymol_cmap, cmin=0, cmax=39, line_w=line_w, ax=ax1)
  475. ims.append([cf.add_text(f"{l}", ax1),img])
  477. ani = animation.ArtistAnimation(fig, ims, blit=True, interval=120)
  478. plt.close()
  479. return ani.to_html5_video()
  481. labels = np.array([f"seed:{x}" for x in range(len(o["traj"]))])
  482. pos = np.array([x["xyz"] for x in o["traj"]])
  483. plddt = np.array([x["plddt"] for x in o["traj"]])
  484. if use_pca and pos.shape[0] > 1:
  485. pos_ca = pos
  486. if o["lengths"] is not None and len(o["lengths"]) > 1:
  487. pos_ca = pos_ca[:,:o["lengths"][0]]
  488. i,j = np.triu_indices(pos_ca.shape[1],k=1)
  489. pos_ca_dm = np.sqrt(np.square(pos_ca[:,None,:,:] - pos_ca[:,:,None]).sum(-1))[:,i,j]
  490. pc = PCA(1).fit_transform(pos_ca_dm)[:,0].argsort()
  491. if cycle:
  492. pc = np.append(pc,pc[1:-1][::-1])
  493. pos = pos[pc]
  494. labels = labels[pc]
  495. plddt = plddt[pc]
  496. HTML(mk_animation(pos, labels, plddt, Ls=o["lengths"], dpi=dpi, color_by_plddt=color_by_plddt))

ESMFold_advanced.ipynb at commit efbf31c, under MIT · at the source

Overview

Authors: Haruka Kanagawa1, Ayaka Kawahara1, Nene Mikawa1, Kana Sugihara1, Momoha Ueda1, Ayano Nitta1, Mizuki Egi1, Reina Oda1, Saori Nonaka2, Hidetoshi Tozaki-Saitoh3, Kosuke Oda4, Hiroshi Nakanishi5
ORCID iDs: Hidetoshi Tozaki-Saitoh, Kosuke Oda, Hiroshi Nakanishi
  1. School of Pharmacy, Yasuda Women’s University, Hiroshima 731-0153, (A.K.); (N.M.); (K.S.); (M.U.); (A.N.); (M.E.); (R.O.)
  2. Laboratory of Molecular Pathobiology, Faculty of Pharmaceutical Sciences, Suzuka University of Medical Science, Mie 513-8670, Japan
  3. Department of Pharmaceutical Sciences, School of Pharmacy at Fukuoka, International University of Health and Welfare, Fukuoka 831-8501, Japan
  4. Department of Pharmacotherapeutics, Faculty of Pharmacy and Pharmaceutical Sciences, Fukuyama University, Hiroshima 729-0292, Japan
  5. Department of Pharmacology, Faculty of Pharmacy, Yasuda Women’s University, Hiroshima 731-0153, Japan
Institutions: Yasuda Women's University (Japan); Suzuka University of Medical Science (Japan); International University of Health and Welfare (Japan); Fukuyama University (Japan)
Journal: International journal of molecular sciences, volume 27, issue 15, article 6614
Dates: received 8 June 2026; accepted 23 July 2026; published online 24 July 2026
Type: Research article · Language: English
License: CC BY
Identifiers: DOI 10.3390/ijms27156614 · PMID 42589276 · PMCID PMC13466725 · OpenAlex W7171422704
Open access: gold, a free copy (OpenAlex)
Status: code verified
Categories: mouse (organism), cellular / molecular (subfield)
Keywords: BV-2 microglia, candidalysin, hydrophobic interactions, interleukin-1β, lipopolysaccharide, nuclear factor-κB, outer membrane lipoproteins, outer membrane vesicles, Prophyromonas gingivalis
MeSH: Interleukin-1beta*, Lipoproteins*, Microglia*, Porphyromonas gingivalis*, Animals, Cell Line, Hydrophobic and Hydrophilic Interactions, Lipopolysaccharides, Mice, NF-kappa B (* major topic)
Topic: Antifungal resistance and susceptibility (Infectious Diseases, Medicine), according to OpenAlex
Funding: JSPS KAKENHI (JP24K09808 (S.N.) and JP24K09787 (K.O.))
Citations: not cited yet (Europe PMC); 36 references in the paper

Abstract

In postmortem Alzheimer’s disease (AD) brains, Porphyromonas gingivalis (Pg), a major periodontal pathogen, and Candida albicans, one of the most common fungal pathogens, have been detected. Although it is important to better understand the effects of their co-infection in the brain for elucidating the pathogenesis of AD, little is known about the neuropathological significance of such co-infection. In the present study, we aimed to elucidate the effects of co-exposure to virulence factors derived from Pg and C. albicans on microglial inflammatory responses. We demonstrated, for the first time, that both candidalysin dissolved in dimethyl sulfoxide (CLd) and water (CLw) significantly suppressed Pg lipopolysaccharide (LPS)-induced interleukin-1β (IL-1β) production by 35–60% and nuclear factor-κB (NF-κB) activation by 20–40%. It should be noted that contaminating Pg outer membrane lipoproteins in Pg LPS were mainly responsible for IL-1β production. To examine the possible hydrophobic interactions between lipoproteins contaminating the Pg LPS preparation and CL, we used 8-anilino-1-naphthalenesulfonic acid sodium salt (ANS-Na), which can be excited to emit fluorescence by binding of hydrophobic molecules. The mean fluorescence intensity of ANS-Na was significantly reduced by approximately 26% following co-treatment with CLw and Pg LPS compared with CLw alone. Furthermore, we generated a mutant form of CL with reduced hydrophobicity (GRAVY index: 1.106 vs. 0.874) while preserving its predicted structural properties. This mutant CLd no longer inhibited Pg LPS-induced IL-1β production. Taken together, these findings indicate that hydrophobic interactions between lipoproteins contaminating the Pg LPS preparation and CL mediate the inhibitory effect of CL on Pg LPS-induced inflammatory responses. The present findings suggest that interactions between polymicrobial virulence factors in the brain may modulate microglia-mediated inflammatory responses during AD progression.

Reproduced under the paper's license (CC BY), from the paper cited above.

Repository

Its files are read in the Code ↔ Paper reader above, with 2 matches between paragraphs and lines of code.

sokrypton/colabfold

License: MIT
State: the link answers, verified on 27 September 2026
Evidence: files inventoried
Commit: efbf31c37cedb38cd09c69c1b991910a9866480e, 19 September 2026
Languages: Python (29), Jupyter (27), Shell (6)
Size: 122 files, 62 scripts
Software Heritage: not archived
Found in: the text, “4.9. AlphaFold Predictions”
Holds: README, license file, environment (Dockerfile, poetry.lock, pyproject.toml), tests, continuous integration, 27 notebooks
Not found: CITATION.cff, documentation
Tools: NumPy (32 files), Matplotlib (24 files), JAX (16 files), PyTorch (8 files), Biopython (7 files), TensorFlow (5 files), SciPy (4 files), PyTorch Lightning (3 files), pandas (3 files), scikit-learn (3 files), PyTorch Geometric (1 file), RDKit (1 file)
Availability: 1 check, the latest on 27 September 2026: the link answers
  • 27 September 2026: the link answers
64 files
At the source: colab.research.google.com/github/sokrypton/ColabFold/blob/main/AlphaFold2.ipynb

Tracing map

Proposed by the machine: these links were found in the paper and verified at the source, without human review. The map will receive a Zenodo DOI once one of the paper's authors has validated it with their ORCID.

What the map holds:

  • 1 repository of the authors' code, each at its verified commit, with its license and how the link was found in the paper;
  • 62 scripts, each with its path and the digest of its content;
  • 2 matches between paragraphs of the paper and lines of the code (method lexical-v1);
  • neither the text of the paper nor the code itself.

Its JSON (tracing-map.json) is deposited on Zenodo with its DOI once the map is validated.

Data

No dataset and no data link were found in the paper.

Data Availability Statement

The original contributions presented in this study are included in the article. Further inquiries can be directed to the corresponding author.

Reproduced under the paper's license (CC BY), from the paper cited above.

Versions

The history of this record: each version stored by the harvester or made by a correction of its authors or of the maintainers of its code, and what changed in its facts. The texts of the paper (its abstract, its availability statements) are not part of it; versions that changed only those are not listed.

Version 1, 27 September 2026: the first record

Recorded: type, language, journal, volume, issue, pages, dates, 12 authors, 9 keywords, 10 MeSH terms, 1 funder, 35 references.

Cite

This paper

Kanagawa, H., Kawahara, A., Mikawa, N., Sugihara, K., Ueda, M., Nitta, A., Egi, M., Oda, R., Nonaka, S., Tozaki-Saitoh, H., Oda, K., & Nakanishi, H. (2026). Candidalysin Inhibits &lt;i&gt;Porphyromonas gingivalis&lt;/i&gt; Lipoprotein-Induced IL-1β Production in BV-2 Microglia via Hydrophobic Microbial Interactions. International journal of molecular sciences, 27(15), 6614. https://doi.org/10.3390/ijms27156614

BibTeX

@article{kanagawa2026candidalysin,
author = {Kanagawa, Haruka and Kawahara, Ayaka and Mikawa, Nene and Sugihara, Kana and Ueda, Momoha and Nitta, Ayano and Egi, Mizuki and Oda, Reina and Nonaka, Saori and Tozaki-Saitoh, Hidetoshi and Oda, Kosuke and Nakanishi, Hiroshi},
title = {{Candidalysin Inhibits \&lt;i\&gt;Porphyromonas gingivalis\&lt;/i\&gt; Lipoprotein-Induced IL-1β Production in BV-2 Microglia via Hydrophobic Microbial Interactions}},
journal = {International journal of molecular sciences},
year = {2026},
month = jul,
volume = {27},
number = {15},
pages = {6614},
publisher = {Multidisciplinary Digital Publishing Institute (MDPI)},
issn = {1422-0067},
doi = {10.3390/ijms27156614},
url = {https://doi.org/10.3390/ijms27156614},
pmid = {42589276},
pmcid = {PMC13466725}
}

RIS

TY - JOUR
AU - Kanagawa, Haruka
AU - Kawahara, Ayaka
AU - Mikawa, Nene
AU - Sugihara, Kana
AU - Ueda, Momoha
AU - Nitta, Ayano
AU - Egi, Mizuki
AU - Oda, Reina
AU - Nonaka, Saori
AU - Tozaki-Saitoh, Hidetoshi
AU - Oda, Kosuke
AU - Nakanishi, Hiroshi
TI - Candidalysin Inhibits &lt;i&gt;Porphyromonas gingivalis&lt;/i&gt; Lipoprotein-Induced IL-1β Production in BV-2 Microglia via Hydrophobic Microbial Interactions
T2 - International journal of molecular sciences
J2 - Int J Mol Sci
PY - 2026
DA - 2026/07/24
VL - 27
IS - 15
SP - 6614
SN - 1422-0067
PB - Multidisciplinary Digital Publishing Institute (MDPI)
DO - 10.3390/ijms27156614
UR - https://doi.org/10.3390/ijms27156614
LA - en
ER -

CSL-JSON

{
"id": "10.3390/ijms27156614",
"type": "article-journal",
"title": "Candidalysin Inhibits &lt;i&gt;Porphyromonas gingivalis&lt;/i&gt; Lipoprotein-Induced IL-1β Production in BV-2 Microglia via Hydrophobic Microbial Interactions",
"container-title": "International journal of molecular sciences",
"author": [
{
"family": "Kanagawa",
"given": "Haruka"
},
{
"family": "Kawahara",
"given": "Ayaka"
},
{
"family": "Mikawa",
"given": "Nene"
},
{
"family": "Sugihara",
"given": "Kana"
},
{
"family": "Ueda",
"given": "Momoha"
},
{
"family": "Nitta",
"given": "Ayano"
},
{
"family": "Egi",
"given": "Mizuki"
},
{
"family": "Oda",
"given": "Reina"
},
{
"family": "Nonaka",
"given": "Saori"
},
{
"family": "Tozaki-Saitoh",
"given": "Hidetoshi"
},
{
"family": "Oda",
"given": "Kosuke"
},
{
"family": "Nakanishi",
"given": "Hiroshi"
}
],
"container-title-short": "Int J Mol Sci",
"volume": "27",
"issue": "15",
"page": "6614",
"DOI": "10.3390/ijms27156614",
"PMID": "42589276",
"PMCID": "PMC13466725",
"ISSN": "1422-0067",
"publisher": "Multidisciplinary Digital Publishing Institute (MDPI)",
"URL": "https://doi.org/10.3390/ijms27156614",
"language": "en",
"issued": {
"date-parts": [
[
2026,
7,
24
]
]
}
}

The tracing map gets a citation of its own once an author has validated it and it has a DOI.

Similar papers

The papers with a page that share the most with this one: the tools found in their code, their categories, datasets, cited references and authors, the rarest counting most.

[1] doi:10.1038/s41586-026-10391-0 [code]
Cell-type-targeted mitochondrial transplantation rescues cell degeneration.
Journal: Nature
In common: RDKit, JAX, PyTorch Lightning, 9 other tools, mouse, cellular / molecular, 1 reference
[2] doi:10.1038/s41598-026-53415-5 [code]
Computational design and immunoinformatics validation of a T cell multi-epitope vaccine targeting glioblastoma stem cells.
Journal: Scientific reports
In common: RDKit, JAX, PyTorch Lightning, 9 other tools, cellular / molecular, 1 reference
[3] doi:10.1038/s42003-026-10957-8 [code]
Brain defence by the extracellular matrix protein Cochlin.
Journal: Communications biology
In common: RDKit, JAX, PyTorch Lightning, 8 other tools, mouse, cellular / molecular
[4] doi:10.1016/j.molcel.2026.07.006 [code]
DeorphaNN: Virtual screening of GPCR peptide agonists using AlphaFold-predicted active-state complexes and deep learning embeddings.
Journal: Molecular cell
In common: JAX, Biopython, PyTorch Geometric, 7 other tools, cellular / molecular, 1 reference
[5] doi:10.1523/eneuro.0362-25.2026 [code]
Similarities between &lt;i&gt;Ciona&lt;/i&gt; Dorsal Motor Ganglion and Vertebrate Cerebellum: Did a Chordate Ancestor Already Show D/V Subdivision within a Hindbrain Precursor?
Journal: eNeuro
In common: PyTorch Lightning, Biopython, PyTorch Geometric, 6 other tools
[6] doi:10.1093/nar/gkag706 [code]
scDifformer: diffusion-based post-training for virtual cell modeling across large-scale single-cell data.
Journal: Nucleic acids research
In common: RDKit, PyTorch Geometric, TensorFlow, 6 other tools
[7] doi:10.1038/s41467-026-71391-2 [code]
Accelerating Leigh syndrome drug discovery through deep learning screening in brain organoids.
Journal: Nature communications
In common: PyTorch Lightning, PyTorch Geometric, TensorFlow, 6 other tools
[8] doi:10.1038/s41467-026-75444-4 [code]
Structural insights enable drug discovery for the neuronal NBCn2 carbonate transporter.
Journal: Nature communications
In common: RDKit, JAX, Biopython, 4 other tools, mouse, cellular / molecular
[9] doi:10.1016/j.xops.2026.101249 [code]
A Disorder-Aware Computational Framework to Identify Structurally Tractable Targets in Proliferative Vitreoretinopathy.
Journal: Ophthalmology science
In common: JAX, Biopython, TensorFlow, 5 other tools
[10] doi:10.1038/s41586-026-10670-w [code]
Zero-shot design of drug-binding proteins via neural iterative selection-expansion.
Journal: Nature
In common: RDKit, PyTorch Geometric, PyTorch, 5 other tools, cellular / molecular

Contribute

The authors of this paper can claim it, correct its record and validate its tracing map, and the maintainers of its code (its owner, or a public member of its organization) correct what it says of their repository; anyone signed in can ask for its removal. Every request goes to OSCR's own machine, which answers it; your account page follows them.

Sign in with ORCID to claim this paper as one of its authors, correct its record or validate its tracing map: when the paper's metadata lists your ORCID iD, you are recognized at once. Maintainers of its code: sign in with GitHub, then claim the repository on your account page.

Request its removal

To ask OSCR to remove this record, the copies of its authors' scripts or its tracing map, use the removal request page: signed in, you say who you are, what to remove and why, then review and confirm the request. Published rules decide every request (how).

Discussion, reproductions, activity

Discussion: questions and error reports about this paper and its code, from signed-in readers and its authors. It opens with sign-in.

Reproductions: reports from readers who ran the authors' code: what they reproduced, with which environment, commit and data. It opens with sign-in.

Activity: what happens around this paper: new versions of its record, its map's validation, discussions and reproductions. It opens with sign-in.