🎨 remove centering

- just does not look good in table of contents

scripts/2_Run_CLASTER.ipynb

@@ -5617,7 +5617,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## <center> Training CLASTER <center>\n",
+    "## Training CLASTER\n",
     "The model can be trained by running the following command:\n",
     "\n",
     "```bash\n",
@@ -5689,7 +5689,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## <center> Testing CLASTER <center>\n",
+    "## Testing CLASTER\n",
     "\n",
     "> _Note:_ \n",
     ">\n",
@@ -5776,7 +5776,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## <center> Predicting _in silico_ perturbed chromatin landscapes <center>\n",
+    "## Predicting _in silico_ perturbed chromatin landscapes\n",
     "\n",
     ">Note: Make sure to create the output folder beforehand: ./runs/perturbation_runs/gene_expression_only_chrom_pure_conv/"
    ]

scripts/3_Data_analysis.ipynb

@@ -958,7 +958,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## <center> Model performance on the baseline prediction task <center>\n",
+    "## Model performance on the baseline prediction task\n",
     "**Figure 1 and related supplementary figures**\n",
     "\n",
     "> Note: ⏰ This took 42 min when analysing all genes and enhancers in chr4:\n",
@@ -1256,7 +1256,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## <center> _In silico_ perturbations <center>\n",
+    "## _In silico_ perturbations\n",
     "\n",
     "**Figure 2 and related supplementary figures**\n",
     "\n"
@@ -1962,7 +1962,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# <center> Attributions <center>\n",
+    "# Attributions\n",
     "\n",
     "Now we will compute the attribution scores for every input position (4,10.001) towards every output (401). This mechanism tells how important is every input position (i.e. a given bin in each input track in the landscape or a given bin in a Micro-C matrix) for the prediction of the EU-seq levels at the position described by a given target node (401 targets). It also tells us the direction of the association.\n",
     "The original paper can be found at:\n",