![\"wedap](\"wedap_logo4.png\"){kind=logo}
\n",
"\n",
" \n",
"
\n",
" ![\"CLI](\"cli2.png\")
\n",
"
\n",
" ![\"GUI\"](\"gui3_postproc.png\")
\n",
"
![\"example](\"example.gif\")
\n",
""
]
},
{
"cell_type": "markdown",
"id": "74f65e65-e98f-469c-ae1f-61336be50caf",
"metadata": {},
"source": [
"---\n",
"Here is an accessible version of the backend code for doing this, I provide it here since you can easily customize it with finer control to make more advanced gifs. "
]
},
{
"cell_type": "code",
"execution_count": 67,
"id": "c90a5dc5",
"metadata": {
"scrolled": true
},
"outputs": [
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "0adaaf35f28c4c919ff4e57a4142fd10",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
" 0%| | 0/18 [00:00"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"# load h5 file into pdist class\n",
"data = wedap.H5_Pdist(\"p53.h5\", data_type=\"average\")\n",
"\n",
"# extract weights\n",
"weights = data.get_all_weights()\n",
"\n",
"# extract data arrays (can be pcoord or any aux data name)\n",
"X = data.get_total_data_array(\"pcoord\", 0)\n",
"Y = data.get_total_data_array(\"pcoord\", 1)\n",
"\n",
"# put X and Y together column wise\n",
"XY = np.hstack((X,Y))\n",
"\n",
"# scale data\n",
"scaler = StandardScaler()\n",
"XY = scaler.fit_transform(XY)\n",
"\n",
"# use 10x less data for easier plotting\n",
"XY = XY[::10,:]\n",
"\n",
"# -ln(W/W(max)) weights\n",
"weights_expanded = -np.log(weights/np.max(weights))[::10]\n",
"\n",
"# cluster pdist using weighted k-means\n",
"clust = KMeans(n_clusters=5).fit(XY, sample_weight=weights_expanded)\n",
"\n",
"# create plot base\n",
"fig, ax = plt.subplots()\n",
"\n",
"# get color labels\n",
"cmap = np.array([\"#377eb8\", \"#ff7f00\", \"#4daf4a\", \"#f781bf\", \"#a65628\"])\n",
"colors = [cmap[label] for label in clust.labels_.astype(int)]\n",
"\n",
"# plot on PCs\n",
"pca = PCA(n_components=2)\n",
"PCs = pca.fit_transform(XY)\n",
"ax.scatter(PCs[:,0], PCs[:,1], c=colors, s=1)\n",
"\n",
"# labels\n",
"ax.set_xlabel(\"PC1\")\n",
"ax.set_ylabel(\"PC2\")\n",
"\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 65,
"id": "a412d8e7",
"metadata": {},
"outputs": [],
"source": [
"# save the resulting figure\n",
"fig.tight_layout()\n",
"fig.savefig(\"km_pca.pdf\")"
]
},
{
"attachments": {},
"cell_type": "markdown",
"id": "bc0e9d84",
"metadata": {},
"source": [
"Now we have a plot along the first two principal components with cluster labels as the colors. Note that this isn't necessarily the most rigorous example, but more of a demonstration of how to use `wedap` with a library like `sklearn`."
]
}
],
"metadata": {
"kernelspec": {
"display_name": "wedap",
"language": "python",
"name": "wedap"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.8.16"
}
},
"nbformat": 4,
"nbformat_minor": 5
}