OSCR

BetaII-spectrin gaps and patches emerge from the patterned assembly of the actin/spectrin membrane skeleton in human motor neuron axons.

Code ↔ Paper

1 match 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 1 match
  1. [1] § Materials and methods › MPS organization assessment › MPS organization assessment by ‘Gollum’ () ↔ ringfinder/ringFinder.py, lines 452 to 594 · score 0.50 · Pearson correlation coefficient, Gollum, tool · Report a mismatch

Paper

Loaded from Europe PMC by your browser; OSCR does not store it: doi.org · Europe PMC

The paper is loaded when this pane is shown.

The authors' code

Python · 676 lines · 29 KB · GPL-3.0 · 1 match

  1. # -*- coding: utf-8 -*-
  2. """
  3. Created on Fri Jul 15 12:25:40 2016
  5. @author: Luciano Masullo, Federico Barabas
  6. """
  8. import os
  9. import time
  10. import math
  11. import numpy as np
  12. from scipy import ndimage as ndi
  13. import tifffile as tiff
  14. from PIL import Image
  16. import pyqtgraph as pg
  17. from pyqtgraph.Qt import QtGui, QtCore
  19. import matplotlib.pyplot as plt
  20. import matplotlib.colors
  21. # from matplotlib import rc
  23. import ringfinder.utils as utils
  24. import ringfinder.tools as tools
  25. import ringfinder.pyqtsubclass as pyqtsub
  27. # rc('text', usetex=True)
  28. # rc('font', **{'family': 'serif', 'serif': ['Computer Modern'], 'size': 15})
  31. class Gollum(QtGui.QMainWindow):
  33. def __init__(self, *args, **kwargs):
  34. super().__init__(*args, **kwargs)
  36. self.i = 0
  37. self.testData = False
  39. self.setWindowTitle('Gollum: the Ring Finder')
  41. self.cwidget = QtGui.QWidget()
  42. self.setCentralWidget(self.cwidget)
  44. menubar = self.menuBar()
  45. fileMenu = menubar.addMenu('&Run')
  46. batchSTORMAct = QtGui.QAction('Analyze batch of STORM images...', self)
  47. batchSTEDAct = QtGui.QAction('Analyze batch of STED images...', self)
  48. batchSTORMAct.triggered.connect(self.batchSTORM)
  49. batchSTEDAct.triggered.connect(self.batchSTED)
  50. fileMenu.addAction(batchSTORMAct)
  51. fileMenu.addAction(batchSTEDAct)
  52. fileMenu.addSeparator()
  54. exitAction = QtGui.QAction(QtGui.QIcon('exit.png'), '&Exit', self)
  55. exitAction.setShortcut('Ctrl+Q')
  56. exitAction.setStatusTip('Exit application')
  57. exitAction.triggered.connect(QtGui.QApplication.closeAllWindows)
  58. fileMenu.addAction(exitAction)
  60. self.folderStatus = QtGui.QLabel('Ready', self)
  61. self.statusBar().addPermanentWidget(self.folderStatus, 1)
  62. self.fileStatus = QtGui.QLabel('Ready', self)
  63. self.statusBar().addPermanentWidget(self.fileStatus)
  65. # Main Widgets' layout
  66. self.mainLayout = QtGui.QGridLayout()
  67. self.cwidget.setLayout(self.mainLayout)
  69. # Image with correlation results
  70. self.corrImgWidget = pg.GraphicsLayoutWidget()
  71. self.corrImgItem = pg.ImageItem()
  72. self.corrVb = self.corrImgWidget.addViewBox(col=0, row=0)
  73. self.corrVb.setAspectLocked(True)
  74. self.corrVb.addItem(self.corrImgItem)
  75. self.corrImgHist = pg.HistogramLUTItem()
  76. self.corrImgHist.gradient.loadPreset('thermal')
  77. self.corrImgHist.setImageItem(self.corrImgItem)
  78. self.corrImgHist.vb.setLimits(yMin=0, yMax=20000)
  79. self.corrImgWidget.addItem(self.corrImgHist)
  80. self.corrResult = pg.ImageItem()
  81. self.corrVb.addItem(self.corrResult)
  83. # Image with ring results
  84. self.ringImgWidget = pg.GraphicsLayoutWidget()
  85. self.ringImgItem = pg.ImageItem()
  86. self.ringVb = self.ringImgWidget.addViewBox(col=0, row=0)
  87. self.ringVb.setAspectLocked(True)
  88. self.ringVb.addItem(self.ringImgItem)
  89. self.ringImgHist = pg.HistogramLUTItem()
  90. self.ringImgHist.gradient.loadPreset('thermal')
  91. self.ringImgHist.setImageItem(self.ringImgItem)
  92. self.ringImgHist.vb.setLimits(yMin=0, yMax=20000)
  93. self.ringImgWidget.addItem(self.ringImgHist)
  94. self.ringResult = pg.ImageItem()
  95. self.ringVb.addItem(self.ringResult)
  97. # Separate frame for loading controls
  98. loadTitle = QtGui.QLabel('<b><u>Load Image</u></b>')
  99. loadTitle.setTextFormat(QtCore.Qt.RichText)
  100. loadTitle.setAlignment(QtCore.Qt.AlignCenter)
  101. loadTitle.setStyleSheet("font-size:14px")
  102. self.STORMPxEdit = QtGui.QLineEdit()
  103. self.magnificationEdit = QtGui.QLineEdit()
  104. self.loadSTORMButton = QtGui.QPushButton('Load')
  105. self.loadSTORMButton.setSizePolicy(QtGui.QSizePolicy.Preferred,
  106. QtGui.QSizePolicy.Expanding)
  107. self.STEDPxEdit = QtGui.QLineEdit()
  108. self.loadSTEDButton = QtGui.QPushButton('Load')
  110. # Ring finding method settings frame
  111. self.intThrLabel = QtGui.QLabel('#sigmas threshold from mean')
  112. self.intThresEdit = QtGui.QLineEdit()
  113. self.sigmaEdit = QtGui.QLineEdit()
  114. self.lineLengthEdit = QtGui.QLineEdit()
  115. self.roiSizeEdit = QtGui.QLineEdit()
  116. self.corrSlider = QtGui.QSlider(QtCore.Qt.Horizontal, self)
  117. self.corrSlider.setMinimum(0)
  118. self.corrSlider.setMaximum(250) # Divide by 1000 to get corr value
  119. self.corrSlider.setValue(200)
  120. self.corrThresEdit = QtGui.QLineEdit()
  121. self.minAreaEdit = QtGui.QLineEdit()
  122. self.corrSlider.valueChanged[int].connect(self.sliderChange)
  123. self.showCorrMapCheck = QtGui.QCheckBox('Show coefficient map', self)
  124. self.thetaStepEdit = QtGui.QLineEdit()
  125. self.deltaThEdit = QtGui.QLineEdit()
  126. self.sinPowerEdit = QtGui.QLineEdit()
  127. self.corrButton = QtGui.QPushButton('Run analysis')
  128. self.corrButton.setCheckable(True)
  129. self.corrButton.setFixedHeight(28)
  130. settingsTitle = QtGui.QLabel('<b><u>Ring finding settings</u></b>')
  131. settingsTitle.setTextFormat(QtCore.Qt.RichText)
  132. settingsTitle.setAlignment(QtCore.Qt.AlignCenter)
  133. settingsTitle.setStyleSheet("font-size:14px")
  134. wvlenLabel = QtGui.QLabel('Rings periodicity [nm]')
  135. self.wvlenEdit = QtGui.QLineEdit()
  136. corrThresLabel = QtGui.QLabel('Discrimination threshold')
  138. # Load settings configuration and then connect the update
  139. try:
  140. tools.loadConfig(self)
  141. except:
  142. tools.saveDefaultConfig()
  143. tools.loadConfig(self)
  144. self.STORMPxEdit.editingFinished.connect(self.updateConfig)
  145. self.magnificationEdit.editingFinished.connect(self.updateConfig)
  146. self.STEDPxEdit.editingFinished.connect(self.updateConfig)
  147. self.roiSizeEdit.editingFinished.connect(self.updateConfig)
  148. self.sigmaEdit.editingFinished.connect(self.updateConfig)
  149. self.intThresEdit.editingFinished.connect(self.updateConfig)
  150. self.lineLengthEdit.editingFinished.connect(self.updateConfig)
  151. self.wvlenEdit.editingFinished.connect(self.updateConfig)
  152. self.sinPowerEdit.editingFinished.connect(self.updateConfig)
  153. self.thetaStepEdit.editingFinished.connect(self.updateConfig)
  154. self.deltaThEdit.editingFinished.connect(self.updateConfig)
  155. self.corrThresEdit.editingFinished.connect(self.updateConfig)
  157. self.buttonWidget = QtGui.QWidget()
  158. buttonsLayout = QtGui.QGridLayout()
  159. self.buttonWidget.setLayout(buttonsLayout)
  160. buttonsLayout.addWidget(loadTitle, 0, 0, 1, 3)
  161. buttonsLayout.addWidget(
  162. QtGui.QLabel('<b>STORM</b> pixel size [nm]'), 1, 0)
  163. buttonsLayout.addWidget(self.STORMPxEdit, 1, 1)
  164. buttonsLayout.addWidget(
  165. QtGui.QLabel('<b>STORM</b> magnification'), 2, 0)
  166. buttonsLayout.addWidget(self.magnificationEdit, 2, 1)
  167. buttonsLayout.addWidget(self.loadSTORMButton, 1, 2, 2, 1)
  168. buttonsLayout.addWidget(
  169. QtGui.QLabel('<b>STED</b> pixel size [nm]'), 3, 0)
  170. buttonsLayout.addWidget(self.STEDPxEdit, 3, 1)
  171. buttonsLayout.addWidget(self.loadSTEDButton, 3, 2)
  173. buttonsLayout.addWidget(settingsTitle, 5, 0, 1, 3)
  174. buttonsLayout.addWidget(wvlenLabel, 6, 0, 1, 2)
  175. buttonsLayout.addWidget(self.wvlenEdit, 6, 2)
  176. buttonsLayout.addWidget(corrThresLabel, 7, 0, 1, 2)
  177. buttonsLayout.addWidget(self.corrThresEdit, 7, 2)
  178. buttonsLayout.addWidget(self.corrSlider, 8, 0, 1, 3)
  179. buttonsLayout.addWidget(self.showCorrMapCheck, 9, 0, 1, 3)
  180. buttonsLayout.addWidget(self.corrButton, 10, 0, 1, 3)
  181. buttonsLayout.setRowMinimumHeight(4, 20)
  182. buttonsLayout.setColumnMinimumWidth(0, 140)
  183. self.buttonWidget.setFixedWidth(270)
  185. # layout of the three widgets
  186. self.mainLayout.addWidget(self.buttonWidget, 1, 0)
  187. corrLabel = QtGui.QLabel('Pearson coefficient')
  188. corrLabel.setAlignment(QtCore.Qt.AlignCenter | QtCore.Qt.AlignVCenter)
  189. self.mainLayout.addWidget(corrLabel, 0, 1)
  190. self.mainLayout.addWidget(self.corrImgWidget, 1, 1, 2, 1)
  191. ringLabel = QtGui.QLabel('Rings')
  192. ringLabel.setAlignment(QtCore.Qt.AlignCenter | QtCore.Qt.AlignVCenter)
  193. self.mainLayout.addWidget(ringLabel, 0, 2)
  194. self.mainLayout.addWidget(self.ringImgWidget, 1, 2, 2, 1)
  195. self.mainLayout.setColumnMinimumWidth(1, 600)
  196. self.mainLayout.setColumnMinimumWidth(2, 600)
  198. self.loadSTORMButton.clicked.connect(self.loadSTORM)
  199. self.loadSTEDButton.clicked.connect(self.loadSTED)
  200. self.sigmaEdit.textChanged.connect(self.updateMasks)
  201. self.corrButton.clicked.connect(self.ringFinder)
  203. # Load sample STED image
  204. folder = os.path.join(os.getcwd(), 'ringfinder')
  205. if os.path.exists(folder):
  206. self.folder = folder
  207. self.loadSTED(os.path.join(folder, 'spectrinSTED.tif'))
  208. else:
  209. self.folder = os.getcwd()
  210. self.loadSTED(os.path.join(os.getcwd(), 'spectrinSTED.tif'))
  212. def updateConfig(self):
  213. tools.saveConfig(self)
  215. def sliderChange(self, value):
  216. self.corrThresEdit.setText(str(np.round(0.001*value, 2)))
  217. self.corrEditChange(str(value/1000))
  219. def corrEditChange(self, text):
  220. self.corrSlider.setValue(1000*float(text))
  221. if self.analyzed:
  222. self.corrThres = float(text)
  223. self.ringsBig = np.nan_to_num(self.localCorrBig) > self.corrThres
  224. self.ringsBig = self.ringsBig.astype(float)
  225. self.ringResult.setImage(np.fliplr(np.transpose(self.ringsBig)))
  227. def loadSTED(self, filename=None):
  228. self.loadImage(np.float(self.STEDPxEdit.text()), 'STED',
  229. filename=filename)
  231. def loadSTORM(self, filename=None):
  232. # The STORM image has black borders because it's not possible to
  233. # localize molecules near the edge of the widefield image.
  234. # Therefore we need to crop those 3px borders before running the
  235. # analysis.
  236. mag = np.float(self.magnificationEdit.text())
  237. load = self.loadImage(np.float(self.STORMPxEdit.text()), 'STORM',
  238. crop=int(3*mag), filename=filename)
  239. if load:
  240. self.corrImgHist.setLevels(0, 3)
  241. self.ringImgHist.setLevels(0, 3)
  243. def loadImage(self, pxSize, tt, crop=0, filename=None):
  245. try:
  246. if not(isinstance(filename, str)):
  247. filetypes = ('Tiff file', '*.tif;*.tiff')
  248. self.filename = utils.getFilename('Load ' + tt + ' image',
  249. [filetypes], self.folder)
  250. else:
  251. self.filename = filename
  253. if self.filename is not None:
  255. self.corrButton.setChecked(False)
  256. self.analyzed = False
  258. self.folder = os.path.split(self.filename)[0]
  259. self.crop = np.int(crop)
  260. self.pxSize = pxSize
  261. self.corrVb.clear()
  262. self.corrResult.clear()
  263. self.ringVb.clear()
  264. self.ringResult.clear()
  266. im = Image.open(self.filename)
  267. self.inputData = np.array(im).astype(np.float64)
  268. self.initShape = self.inputData.shape
  269. bound = (np.array(self.initShape) - self.crop).astype(np.int)
  270. self.inputData = self.inputData[self.crop:bound[0],
  271. self.crop:bound[1]]
  272. self.shape = self.inputData.shape
  274. # We need 1um n-sized subimages
  275. self.subimgPxSize = int(1000/self.pxSize)
  276. self.n = (np.array(self.shape)/self.subimgPxSize).astype(int)
  278. # If n*subimgPxSize < shape, we crop the image
  279. self.remanent = np.array(self.shape) - self.n*self.subimgPxSize
  280. self.inputData = self.inputData[:self.n[0]*self.subimgPxSize,
  281. :self.n[1]*self.subimgPxSize]
  282. self.shape = self.inputData.shape
  284. self.nblocks = np.array(self.inputData.shape)/self.n
  285. self.blocksInput = tools.blockshaped(self.inputData,
  286. *self.nblocks)
  288. self.updateMasks()
  289. self.corrVb.addItem(self.corrImgItem)
  290. self.ringVb.addItem(self.ringImgItem)
  291. showIm = np.fliplr(np.transpose(self.inputData))
  292. self.corrImgItem.setImage(showIm)
  293. self.ringImgItem.setImage(showIm)
  295. self.grid = pyqtsub.Grid(self.corrVb, self.shape, self.n)
  297. self.corrVb.setLimits(xMin=-0.05*self.shape[0],
  298. xMax=1.05*self.shape[0], minXRange=4,
  299. yMin=-0.05*self.shape[1],
  300. yMax=1.05*self.shape[1], minYRange=4)
  301. self.ringVb.setLimits(xMin=-0.05*self.shape[0],
  302. xMax=1.05*self.shape[0], minXRange=4,
  303. yMin=-0.05*self.shape[1],
  304. yMax=1.05*self.shape[1], minYRange=4)
  306. self.dataMean = np.mean(self.inputData)
  307. self.dataStd = np.std(self.inputData)
  309. self.corrVb.addItem(self.corrResult)
  310. self.ringVb.addItem(self.ringResult)
  312. return True
  314. else:
  315. return False
  317. except OSError:
  318. self.fileStatus.setText('No file selected!')
  320. def updateMasks(self):
  321. """Binarization of image. """
  323. self.gaussSigma = np.float(self.sigmaEdit.text())/self.pxSize
  324. thr = np.float(self.intThresEdit.text())
  326. if self.testData:
  327. self.blocksInputS = [ndi.gaussian_filter(b, self.gaussSigma)
  328. for b in self.blocksInput]
  329. self.blocksInputS = np.array(self.blocksInputS)
  330. self.meanS = np.mean(self.blocksInputS, (1, 2))
  331. self.stdS = np.std(self.blocksInputS, (1, 2))
  332. thresholds = self.meanS + thr*self.stdS
  333. thresholds = thresholds.reshape(np.prod(self.n), 1, 1)
  334. mask = self.blocksInputS < thresholds
  335. self.blocksMask = np.array([bI < np.mean(bI) + thr*np.std(bI)
  336. for bI in self.blocksInputS])
  338. self.mask = tools.unblockshaped(mask, *self.inputData.shape)
  339. self.inputDataS = tools.unblockshaped(self.blocksInputS,
  340. *self.shape)
  341. else:
  342. self.inputDataS = ndi.gaussian_filter(self.inputData,
  343. self.gaussSigma)
  344. self.blocksInputS = tools.blockshaped(self.inputDataS,
  345. *self.nblocks)
  346. self.meanS = np.mean(self.inputDataS)
  347. self.stdS = np.std(self.inputDataS)
  348. self.mask = self.inputDataS < self.meanS + thr*self.stdS
  349. self.blocksMask = tools.blockshaped(self.mask, *self.nblocks)
  351. self.showImS = np.fliplr(np.transpose(self.inputDataS))
  352. self.showMask = np.fliplr(np.transpose(self.mask))
  354. def ringFinder(self, show=True, batch=False):
  355. """RingFinder handles the input data, and then evaluates every subimg
  356. using the given algorithm which decides if there are rings or not.
  357. Subsequently gives the output data and plots it"""
  359. if self.corrButton.isChecked() or batch:
  361. self.corrResult.clear()
  362. self.ringResult.clear()
  364. # for each subimg, we apply the correlation method for ring finding
  365. intThr = np.float(self.intThresEdit.text())
  366. minLen = np.float(self.lineLengthEdit.text())/self.pxSize
  367. thetaStep = np.float(self.thetaStepEdit.text())
  368. deltaTh = np.float(self.deltaThEdit.text())
  369. wvlen = np.float(self.wvlenEdit.text())/self.pxSize
  370. sinPow = np.float(self.sinPowerEdit.text())
  371. cArgs = minLen, thetaStep, deltaTh, wvlen, sinPow
  373. # Single-core code
  374. self.localCorr = np.zeros(len(self.blocksInput))
  375. thres = self.meanS + intThr*self.stdS
  376. if not(self.testData):
  377. thres = thres*np.ones(self.blocksInput.shape)
  379. for i in np.arange(len(self.blocksInput)):
  380. block = self.blocksInput[i]
  381. blockS = self.blocksInputS[i]
  382. mask = self.blocksMask[i]
  383. # Block may be excluded from the analysis for two reasons.
  384. # Firstly, because the intensity for all its pixels may be
  385. # too low. Secondly, because the part of the block that
  386. # belongs toa neuron may be below an arbitrary 20% of the
  387. # block. We apply intensity threshold to smoothed data so we
  388. # don't catch tiny bright spots outside neurons
  389. neuronFrac = 1 - np.sum(mask)/np.size(mask)
  390. areaThres = 0.01*float(self.minAreaEdit.text())
  391. if np.any(blockS > thres[i]) and neuronFrac > areaThres:
  392. output = tools.corrMethod(block, mask, *cArgs)
  393. angle, corrTheta, corrMax, theta, phase = output
  394. # Store results
  395. self.localCorr[i] = corrMax
  396. else:
  397. self.localCorr[i] = np.nan
  399. self.localCorr = self.localCorr.reshape(*self.n)
  400. self.updateGUI(self.localCorr)
  402. else:
  403. self.corrResult.clear()
  404. self.ringResult.clear()
  406. def updateGUI(self, localCorr):
  408. self.analyzed = True
  409. self.localCorr = localCorr
  411. # code for visualization of the output
  412. mag = np.array(self.inputData.shape)/self.n
  413. self.localCorrBig = np.repeat(self.localCorr, mag[0], 0)
  414. self.localCorrBig = np.repeat(self.localCorrBig, mag[1], 1)
  415. showIm = 100*np.fliplr(np.transpose(self.localCorrBig))
  416. self.corrResult.setImage(np.nan_to_num(showIm))
  417. self.corrResult.setZValue(10) # make sure this image is on top
  418. self.corrResult.setOpacity(0.5)
  420. self.corrThres = float(self.corrThresEdit.text())
  421. self.ringsBig = np.nan_to_num(self.localCorrBig) > self.corrThres
  422. self.ringsBig = self.ringsBig.astype(float)
  423. self.ringResult.setImage(np.fliplr(np.transpose(self.ringsBig)))
  424. self.ringResult.setZValue(10) # make sure this image is on top
  425. self.ringResult.setOpacity(0.5)
  427. if self.showCorrMapCheck.isChecked():
  428. plt.figure(figsize=(10, 8))
  429. data = self.localCorr.reshape(*self.n)
  430. data = np.flipud(data)
  431. maskedData = np.ma.array(data, mask=np.isnan(data))
  433. mx = np.max(maskedData)
  434. mn = np.min(maskedData)
  435. mp = (self.corrThres - mn)/(mx - mn)
  436. mp = np.min((mp, 1))
  437. mp = np.max((mp, 0))
  439. cmap = shiftedColorMap(matplotlib.cm.PuOr, midpoint=mp,
  440. name='shifted')
  441. heatmap = plt.pcolor(maskedData, cmap=cmap)
  442. for y in range(data.shape[0]):
  443. for x in range(data.shape[1]):
  444. plt.text(x + 0.5, y + 0.5, '%.2f' % data[y, x],
  445. horizontalalignment='center',
  446. verticalalignment='center',)
  447. plt.colorbar(heatmap)
  448. plt.gca().set_xticklabels([])
  449. plt.gca().set_yticklabels([])
  450. plt.show()
  452. def batch(self, function, tech):
  453. try:
  454. filenames = utils.getFilenames('Load ' + tech + ' images',
  455. [('Tiff file', '*.tif;*.tiff')],
  456. self.folder)
  457. nfiles = len(filenames)
  458. function(filenames[0])
  459. corrArray = np.zeros((nfiles, self.n[0], self.n[1]))
  461. # Expand correlation array so it matches data shape
  462. corrExp = np.empty((nfiles, self.initShape[0], self.initShape[1]),
  463. dtype=np.single)
  464. corrExp[:] = np.nan
  465. ringsExp = np.empty((nfiles, self.initShape[0], self.initShape[1]),
  466. dtype=np.single)
  467. ringsExp[:] = np.nan
  469. path = os.path.split(filenames[0])[0]
  470. folder = os.path.split(path)[1]
  471. self.folderStatus.setText('Processing folder ' + path)
  472. print('Processing folder', path)
  473. t0 = time.time()
  475. # Make results directory if it doesn't exist
  476. resultsDir = os.path.join(path, 'results')
  477. if not os.path.exists(resultsDir):
  478. os.makedirs(resultsDir)
  479. resNames = [utils.insertFolder(p, 'results') for p in filenames]
  481. for i in np.arange(nfiles):
  482. print(os.path.split(filenames[i])[1])
  483. self.fileStatus.setText(os.path.split(filenames[i])[1])
  484. function(filenames[i])
  485. self.ringFinder(False, batch=True)
  486. corrArray[i] = self.localCorr
  488. bound = (np.array(self.initShape) - self.crop).astype(np.int)
  489. edge = bound - self.remanent
  490. corrExp[i, self.crop:edge[0],
  491. self.crop:edge[1]] = self.localCorrBig
  493. # Save correlation values array
  494. corrName = utils.insertSuffix(resNames[i], '_correlation')
  495. tiff.imsave(corrName, corrExp[i], software='Gollum',
  496. imagej=True,
  497. resolution=(1000/self.pxSize, 1000/self.pxSize),
  498. metadata={'spacing': 1, 'unit': 'um'})
  500. # Saving ring images
  501. ringsExp[corrExp < self.corrThres] = 0
  502. ringsExp[corrExp >= self.corrThres] = 1
  503. for i in np.arange(nfiles):
  504. # Save correlation values array
  505. ringName = utils.insertSuffix(resNames[i], '_rings')
  506. tiff.imsave(ringName, ringsExp[i], software='Gollum',
  507. imagej=True,
  508. resolution=(1000/self.pxSize, 1000/self.pxSize),
  509. metadata={'spacing': 1, 'unit': 'um'})
  511. # save configuration file in the results folder
  512. tools.saveConfig(self, os.path.join(resultsDir, 'config'))
  514. # plot histogram of the correlation values
  515. hrange = (np.min(np.nan_to_num(corrArray)),
  516. np.max(np.nan_to_num(corrArray)))
  517. y, x, _ = plt.hist(corrArray.flatten(), bins=20, range=hrange)
  518. x = (x[1:] + x[:-1])/2
  520. # Save data array as txt
  521. corrArrayFlat = corrArray.flatten()
  522. validCorr = corrArrayFlat[~np.isnan(corrArrayFlat)]
  523. validArr = np.repeat(np.arange(nfiles), np.prod(self.n))
  524. validArr = validArr[~np.isnan(corrArrayFlat)]
  525. valuesTxt = os.path.join(resultsDir, folder + 'corr_values.txt')
  526. corrByN = np.stack((validCorr, validArr), 1)
  527. np.savetxt(valuesTxt, corrByN, fmt='%f\t%i')
  529. groupedCorr = [validCorr[np.where(validArr == i)]
  530. for i in np.arange(nfiles)]
  531. groupedCorr = [x for x in groupedCorr if len(x) > 0]
  532. nfiles = len(groupedCorr)
  533. meanCorrs = [np.mean(d) for d in groupedCorr]
  535. validCorrRing = validCorr[np.where(validCorr > self.corrThres)]
  536. validArrRing = validArr[np.where(validCorr > self.corrThres)]
  537. groupedCorrRing = [validCorrRing[np.where(validArrRing == i)]
  538. for i in np.arange(nfiles)]
  539. ringFracs = [len(groupedCorrRing[i])/len(groupedCorr[i])
  540. for i in np.arange(nfiles)]
  541. meanRingCorrs = np.array([np.mean(d) for d in groupedCorrRing])
  542. meanRingCorrs = meanRingCorrs[~np.isnan(meanRingCorrs)]
  544. n = corrArray.size - np.count_nonzero(np.isnan(corrArray))
  545. nring = np.sum(validCorr > self.corrThres)
  546. validRingCorr = validCorr[validCorr > self.corrThres]
  547. ringFrac = nring/n
  549. # Err estimation: stat err (binomial distribution, p=ringFrac)
  550. statVar = ringFrac*(1 - ringFrac)/n
  551. fracStd = math.sqrt(statVar)
  553. statCorrVar = np.var(validCorr)/n
  554. corrStd = math.sqrt(statCorrVar)
  556. statRingCorrVar = np.var(validRingCorr)/nring
  557. ringCorrStd = math.sqrt(statRingCorrVar)
  559. # Plotting
  560. plt.style.use('ggplot')
  561. plt.figure(figsize=(10, 7.5))
  562. plt.bar(x, y, align='center', width=(x[1] - x[0]), color="#3F5D7D")
  563. plt.plot((self.corrThres, self.corrThres), (0, np.max(y)), '--',
  564. color='r', linewidth=2)
  565. text = ('Pearson coefficient threshold = {0:.2f} \n'
  566. 'n = {1}; nrings = {2} \n'
  567. 'PSS fraction = {3:.2f} $\pm$ {4:.2f} \n'
  568. 'mean coefficient = {5:.3f} $\pm$ {6:.3f}\n'
  569. 'mean ring coefficient = {7:.3f} $\pm$ {8:.3f}')
  570. text = text.format(self.corrThres, n, nring,
  571. np.mean(ringFracs), fracStd,
  572. np.mean(meanCorrs), corrStd,
  573. np.mean(meanRingCorrs), ringCorrStd)
  574. plt.text(0.75*plt.axis()[1], 0.83*plt.axis()[3], text,
  575. horizontalalignment='center', verticalalignment='center',
  576. bbox=dict(facecolor='white'), fontsize=20)
  577. plt.xlabel('Pearson correlation coefficient', fontsize=35)
  578. plt.tick_params(axis='both', labelsize=25)
  579. plt.grid()
  580. plt.tight_layout()
  581. plt.savefig(os.path.join(resultsDir, folder + 'corr_hist.pdf'),
  582. dpi=300)
  583. plt.savefig(os.path.join(resultsDir, folder + 'corr_hist.png'),
  584. dpi=300)
  585. plt.close()
  587. folder = os.path.split(path)[1]
  588. text = 'Folder ' + folder + ' done in {0:.0f} seconds'
  589. print(text.format(time.time() - t0))
  590. self.folderStatus.setText(text.format(time.time() - t0))
  591. self.fileStatus.setText(' ')
  593. except IndexError:
  594. self.fileStatus.setText('No file selected!')
  596. def batchSTORM(self):
  597. self.batch(self.loadSTORM, 'STORM')
  599. def batchSTED(self):
  600. self.batch(self.loadSTED, 'STED')
  603. def shiftedColorMap(cmap, start=0, midpoint=0.5, stop=1.0, name='shiftedcmap'):
  604. '''
  605. Function to offset the "center" of a colormap. Useful for
  606. data with a negative min and positive max and you want the
  607. middle of the colormap's dynamic range to be at zero
  609. Input
  610. -----
  611. cmap : The matplotlib colormap to be altered
  612. start : Offset from lowest point in the colormap's range.
  613. Defaults to 0.0 (no lower ofset). Should be between
  614. 0.0 and `midpoint`.
  615. midpoint : The new center of the colormap. Defaults to
  616. 0.5 (no shift). Should be between 0.0 and 1.0. In
  617. general, this should be 1 - vmax/(vmax + abs(vmin))
  618. For example if your data range from -15.0 to +5.0 and
  619. you want the center of the colormap at 0.0, `midpoint`
  620. should be set to 1 - 5/(5 + 15)) or 0.75
  621. stop : Offset from highets point in the colormap's range.
  622. Defaults to 1.0 (no upper ofset). Should be between
  623. `midpoint` and 1.0.
  624. http://stackoverflow.com/questions/7404116/
  625. defining-the-midpoint-of-a-colormap-in-matplotlib
  626. '''
  628. if midpoint == stop:
  629. newcmap = truncate_colormap(cmap, 0, 0.5)
  631. elif midpoint == start:
  632. newcmap = truncate_colormap(cmap, 0.5, 1)
  634. else:
  635. cdict = {
  636. 'red': [],
  637. 'green': [],
  638. 'blue': [],
  639. 'alpha': []
  640. }
  642. # regular index to compute the colors
  643. reg_index = np.linspace(start, stop, 257)
  645. # shifted index to match the data
  646. shift_index = np.hstack([
  647. np.linspace(0.0, midpoint, 128, endpoint=False),
  648. np.linspace(midpoint, 1.0, 129, endpoint=True)
  649. ])
  651. for ri, si in zip(reg_index, shift_index):
  652. r, g, b, a = cmap(ri)
  654. cdict['red'].append((si, r, r))
  655. cdict['green'].append((si, g, g))
  656. cdict['blue'].append((si, b, b))
  657. cdict['alpha'].append((si, a, a))
  659. newcmap = matplotlib.colors.LinearSegmentedColormap(name, cdict)
  661. plt.register_cmap(cmap=newcmap)
  663. return newcmap
  666. def truncate_colormap(cmap, minval=0.0, maxval=1.0, n=256):
  667. new_cmap = matplotlib.colors.LinearSegmentedColormap.from_list(
  668. 'trunc({n},{a:.2f},{b:.2f})'.format(n=cmap.name, a=minval, b=maxval),
  669. cmap(np.linspace(minval, maxval, n)))
  670. return new_cmap
  672. if __name__ == '__main__':
  673. app = QtGui.QApplication([])
  674. win = Gollum()
  675. win.show()
  676. app.exec_()

ringFinder.py at commit b97868c, under GPL-3.0 · at the source

Overview

Authors: Nahir Guadalupe Gazal1, Maria Jose Castellanos-Montiel2, Guillermina Bruno1, Anna Kristina Franco-Flores2, Sarah Lépine2,3, Lale Gursu2, Ghazal Haghi2, Gilles Maussion2, Wolfgang E Reintsch2, Fernando D Stefani4,5, Agustín Anastasía1,6, Mariano Bisbal1,6,7, Ezequiel Axel Gorostiza8, Thomas M Durcan2, Nicolás Unsain1,6,7
ORCID iDs: Gilles Maussion, Mariano Bisbal, Thomas M Durcan, Nicolás Unsain
  1. Instituto de Investigación Médica Mercedes y Martín Ferreyra (INIMEC), Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Universidad Nacional de Córdoba Córdoba Argentina
  2. Early Drug Discovery Unit (EDDU), The Neuro-Montreal Neurological Institute and Hospital, Department of Neurology and Neurosurgery, McGill University Montreal Canada
  3. Faculty of Medicine and Health Sciences, McGill University Montreal Canada
  4. Centro de Investigaciones en Bionanociencias (CIBION), Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET) Ciudad Autónoma de Buenos Aires Argentina
  5. Departamento de Física, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires Ciudad Autónoma de Buenos Aires Argentina
  6. Instituto Universitario de Ciencias Biomédicas de Córdoba (IUCBC) Córdoba Argentina
  7. Facultad de Ciencias Exactas, Físicas y Naturales, Universidad Nacional de Córdoba Córdoba Argentina
  8. Institute of Zoology, Biocenter Cologne, University of Cologne Cologne Germany
Institutions: Universidad Nacional de Córdoba (Argentina); McGill University (Canada); Consejo Nacional de Investigaciones Científicas y Técnicas (Argentina); Universidad de Buenos Aires (Argentina); Instituto Universitario de Ciencias Biomédicas de Córdoba (Argentina); University of Cologne (Germany)
Journal: eLife, volume 14, article RP108021
Dates: published online 8 September 2026
Type: Research article · Language: English
License: CC BY
Identifiers: DOI 10.7554/elife.108021 · PMID 42708447 · PMCID PMC13553066 · OpenAlex W4414281024
Open access: gold, a free copy (OpenAlex)
Status: code verified (how the code is found and verified)
Categories: human (organism), mouse (organism), cellular / molecular (subfield)
Methods: Statistics, Connectivity, fMRI & imaging
Keywords: spectrin, axon, actin, motor neurons, MPS, cytoskeleton, iPSCs, staurosporine, Human, Mouse
MeSH: Actins*, Axons*, Cytoskeleton*, Motor Neurons*, Spectrin*, Humans, Induced Pluripotent Stem Cells (* major topic)
Journal subjects: Cell Biology, Neuroscience
Topic: Cellular Mechanics and Interactions (Cell Biology, Biochemistry, Genetics and Molecular Biology), according to OpenAlex
Funding: Fondo para la Investigación Científica y Tecnológica (ANPCyT-PICT-2021-GRF-II-00048); International Brain Research Organization (IBRO-Collaboration Grants 2021); ALS Society of Canada
Citations: not cited yet (Europe PMC); 61 references in the paper

Abstract

The actin/spectrin membrane-associated periodic skeleton (MPS) is a cytoskeletal structure that supports axonal integrity and function. Lower spinal motor neurons (MNs) are characterized by exceptionally long axons and are particularly susceptible to degeneration in a wide range of hereditary neuromuscular disorders, including amyotrophic lateral sclerosis. Using confocal and super-resolution imaging, we characterized the spatial distribution of βII-spectrin and the assembly pattern of the MPS in human MN axons derived from induced pluripotent stem cells. We discovered a striking gap-and-patch pattern in the medial axon, where sharply demarcated βII-spectrin gaps alternate with patches containing a well-organized MPS. The pattern is acutely induced by the kinase inhibitor staurosporine and pharmacological inhibition of actin polymerization prevents patch formation, indicating a requirement for actin nucleation in MPS assembly. Our data supports a model in which spectrin incorporation into nascent MPS patches depletes neighboring regions, producing long-range gaps-and-patches patterns.

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 1 match between paragraphs and lines of code.

cibion-conicet/Gollum

License: GPL-3.0
State: the link answers, verified on 26 September 2026
Evidence: files inventoried
Commit: b97868cff590ccdf6df0dfe61a939f85cf83a47d, 7 February 2018
Languages: Python (13)
Size: 24 files, 13 scripts
Software Heritage: not archived
Found in: the text, “MPS organization assessment by ‘Gollum’ ( Baraba”
Holds: README, license file, environment (requirements.txt)
Not found: CITATION.cff, tests, continuous integration, documentation
Tools: NumPy (7 files), SciPy (4 files), Matplotlib (3 files), Pillow (3 files), tifffile (3 files), scikit-image (1 file)
Availability: 1 check, the latest on 26 September 2026: the link answers
  • 26 September 2026: the link answers
15 files
At the source: github.com/cibion-conicet/Gollum

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;
  • 13 scripts, each with its path and the digest of its content;
  • 1 match 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

All data generated or analyzed during this study are included in this published article. The raw data used in this article has been deposited in the following domain: http://hdl.handle.net/11336/294857. The cell lines utilized in this work were created by the The Early Drug Discovery Unit (EDDU) at The Neuro (Montreal Neurological Institute and Hospital, McGill University). The cell lines are deposited in The Neuro C-BIG repository and are publicly available (https://www.neuro-edduportal.com/ipsc-catalogue).

The following dataset was generated:

AnastasiaA BisbalM UnsainN 2026Raw microscopy images from cultured human-derived motor neurons stained for b2-spectrin and other cytoskeletal proteins and registered by confocal and STED microsocypy: the whole set was used to produce the publication entitled "BetaII-Spectrin Gaps and Patches Emerge from the Patterned Assembly of the Actin/Spectrin Membrane Skeleton in Human Motor Neuron AxonsHandle11336/294857

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, pages, dates, 15 authors, 10 keywords, 7 MeSH terms, 3 funders, 60 references.

Cite

This paper

Gazal, N. G., Castellanos-Montiel, M. J., Bruno, G., Franco-Flores, A. K., Lépine, S., Gursu, L., Haghi, G., Maussion, G., Reintsch, W. E., Stefani, F. D., Anastasía, A., Bisbal, M., Gorostiza, E. A., Durcan, T. M., & Unsain, N. (2026). BetaII-spectrin gaps and patches emerge from the patterned assembly of the actin/spectrin membrane skeleton in human motor neuron axons. eLife, 14, RP108021. https://doi.org/10.7554/elife.108021

BibTeX

@article{gazal2026betaii,
author = {Gazal, Nahir Guadalupe and Castellanos-Montiel, Maria Jose and Bruno, Guillermina and Franco-Flores, Anna Kristina and Lépine, Sarah and Gursu, Lale and Haghi, Ghazal and Maussion, Gilles and Reintsch, Wolfgang E and Stefani, Fernando D and Anastasía, Agustín and Bisbal, Mariano and Gorostiza, Ezequiel Axel and Durcan, Thomas M and Unsain, Nicolás},
title = {{BetaII-spectrin gaps and patches emerge from the patterned assembly of the actin/spectrin membrane skeleton in human motor neuron axons}},
journal = {eLife},
year = {2026},
month = sep,
volume = {14},
pages = {RP108021},
publisher = {eLife Sciences Publications, Ltd},
issn = {2050-084X},
doi = {10.7554/elife.108021},
url = {https://doi.org/10.7554/elife.108021},
pmid = {42708447},
pmcid = {PMC13553066}
}

RIS

TY - JOUR
AU - Gazal, Nahir Guadalupe
AU - Castellanos-Montiel, Maria Jose
AU - Bruno, Guillermina
AU - Franco-Flores, Anna Kristina
AU - Lépine, Sarah
AU - Gursu, Lale
AU - Haghi, Ghazal
AU - Maussion, Gilles
AU - Reintsch, Wolfgang E
AU - Stefani, Fernando D
AU - Anastasía, Agustín
AU - Bisbal, Mariano
AU - Gorostiza, Ezequiel Axel
AU - Durcan, Thomas M
AU - Unsain, Nicolás
TI - BetaII-spectrin gaps and patches emerge from the patterned assembly of the actin/spectrin membrane skeleton in human motor neuron axons
T2 - eLife
J2 - eLife
PY - 2026
DA - 2026/09/08
VL - 14
SP - RP108021
SN - 2050-084X
PB - eLife Sciences Publications, Ltd
DO - 10.7554/elife.108021
UR - https://doi.org/10.7554/elife.108021
LA - en
ER -

CSL-JSON

{
"id": "10.7554/elife.108021",
"type": "article-journal",
"title": "BetaII-spectrin gaps and patches emerge from the patterned assembly of the actin/spectrin membrane skeleton in human motor neuron axons",
"container-title": "eLife",
"author": [
{
"family": "Gazal",
"given": "Nahir Guadalupe"
},
{
"family": "Castellanos-Montiel",
"given": "Maria Jose"
},
{
"family": "Bruno",
"given": "Guillermina"
},
{
"family": "Franco-Flores",
"given": "Anna Kristina"
},
{
"family": "Lépine",
"given": "Sarah"
},
{
"family": "Gursu",
"given": "Lale"
},
{
"family": "Haghi",
"given": "Ghazal"
},
{
"family": "Maussion",
"given": "Gilles"
},
{
"family": "Reintsch",
"given": "Wolfgang E"
},
{
"family": "Stefani",
"given": "Fernando D"
},
{
"family": "Anastasía",
"given": "Agustín"
},
{
"family": "Bisbal",
"given": "Mariano"
},
{
"family": "Gorostiza",
"given": "Ezequiel Axel"
},
{
"family": "Durcan",
"given": "Thomas M"
},
{
"family": "Unsain",
"given": "Nicolás"
}
],
"container-title-short": "eLife",
"volume": "14",
"page": "RP108021",
"DOI": "10.7554/elife.108021",
"PMID": "42708447",
"PMCID": "PMC13553066",
"ISSN": "2050-084X",
"publisher": "eLife Sciences Publications, Ltd",
"URL": "https://doi.org/10.7554/elife.108021",
"language": "en",
"issued": {
"date-parts": [
[
2026,
9,
8
]
]
}
}

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/s41467-022-30720-x [code]
Proteomic and functional analyses of the periodic membrane skeleton in neurons.
Journal: Nature communications
In common: tifffile, Pillow, SciPy, 2 other tools, cellular / molecular, 16 references
[2] doi:10.1126/sciadv.aeb0803 [code]
Membrane-associated periodic skeleton regulates major forms of endocytosis in neurons through a signaling-driven positive feedback loop.
Journal: Science advances
In common: tifffile, Pillow, SciPy, 2 other tools, mouse, cellular / molecular, 13 references
[3] doi:10.1016/j.isci.2025.114454
Spectrin condensates provide a nidus for assembling the axonal membrane-associated periodic skeleton.
Journal: iScience
In common: cellular / molecular, 16 references
[4] doi:10.1126/sciadv.adt3724
Paralemmin-1 controls the nanoarchitecture of the neuronal submembrane cytoskeleton.
Journal: Science advances
In common: mouse, cellular / molecular, 11 references
[5] doi:10.1016/j.isci.2024.109166 [code]
Homozygous ALS-linked mutations in TARDBP/TDP-43 lead to hypoactivity and synaptic abnormalities in human iPSC-derived motor neurons.
Journal: iScience
In common: SciPy, Matplotlib, NumPy, cellular / molecular, 5 references, author Thomas M Durcan
[6] doi:10.3390/ijms241914926 [code]
Transcriptional Dysregulation and Impaired Neuronal Activity in <i>FMR1</i> Knock-Out and Fragile X Patients' iPSC-Derived Models.
Journal: International journal of molecular sciences
In common: mouse, cellular / molecular, 2 references, 2 authors
[7] doi:10.1016/j.cub.2026.03.053
External mechanical force drives axonal cytoskeletal rearrangements.
Journal: Current biology : CB
In common: cellular / molecular, 8 references
[8] doi:10.3390/cells12040545 [code]
An Optimized Workflow to Generate and Characterize iPSC-Derived Motor Neuron (MN) Spheroids.
Journal: Cells
In common: cellular / molecular, 4 references, author Thomas M Durcan
[9] doi:10.1073/pnas.2310120121 [code]
An ankyrin G-binding motif mediates TRAAK periodic localization at axon initial segments of hippocampal pyramidal neurons.
Journal: Proceedings of the National Academy of Sciences of the United States of America
In common: scikit-image, SciPy, Matplotlib, 1 other tool, cellular / molecular, 4 references
[10] doi:10.1038/s41467-025-60709-1 [code]
Structural Repetition Detector for multi-scale quantitative mapping of molecular complexes through microscopy.
Journal: Nature communications
In common: tifffile, SciPy, Matplotlib, 1 other tool, cellular / molecular, 4 references

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.

Add the badge to the code's README

OSCR: the badge of this paper

If you are an author or a maintainer of this paper's code, you can add this badge. It says what OSCR knows of the paper (its code, its validated tracing map) and links the code to this page. Paste one of these snippets under the README's title. Keep the two marks around it, so that the badge can be replaced in place and is never added twice. OSCR changes nothing in your repository.

Markdown
<!-- oscr-badge:start -->
[![OSCR](https://openscicode.org/badge/doi/10.7554/elife.108021.svg)](https://openscicode.org/badge/doi/10.7554/elife.108021)
<!-- oscr-badge:end -->
reStructuredText
.. oscr-badge:start

.. image:: https://openscicode.org/badge/doi/10.7554/elife.108021.svg
   :target: https://openscicode.org/badge/doi/10.7554/elife.108021
   :alt: OSCR

.. oscr-badge:end
HTML
<!-- oscr-badge:start -->
<a href="https://openscicode.org/badge/doi/10.7554/elife.108021"><img src="https://openscicode.org/badge/doi/10.7554/elife.108021.svg" alt="OSCR"></a>
<!-- oscr-badge:end -->

On GitHub, open the README, choose Edit (the pencil), paste the snippet under the title, then "Commit changes".

If you choose to, you can have the badge kept up to date for you, with oscr badge --write from the command line or with the GitHub Action in your repository's workflows. Every other form and option (a card, AsciiDoc, LaTeX, shields.io) is on the badge 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: code errors and mismatches between the code and the paper about this paper and its code, from readers and its authors, beside the paper and its tracing map.

Report a mismatch between the code and the paper · Report a code error (signed in)

Reproductions: reports from readers who ran the authors' code and did not get the paper's result, what they ran, with which environment, commit and data.

Report a reproduction failure (signed in)

Versions of its code: the releases of its code tied to a version of this paper (the preprint, the accepted manuscript, the version of record…), each with the tracing map versioned with it.

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