From ec21a244298dec9de00570a6a2aaf2ab39f4787a Mon Sep 17 00:00:00 2001 From: Luis Pedro Coelho Date: Sun, 21 Jan 2024 19:36:16 +0100 Subject: [PATCH] ENH Add four more features Resize everything so that it fits --- .../07_c_AMP_features_comparison.py | 29 ++++++++++++------- 1 file changed, 18 insertions(+), 11 deletions(-) diff --git a/Manuscript_Analysis/07_c_AMP_features_comparison.py b/Manuscript_Analysis/07_c_AMP_features_comparison.py index 0036308..99e4f8b 100644 --- a/Manuscript_Analysis/07_c_AMP_features_comparison.py +++ b/Manuscript_Analysis/07_c_AMP_features_comparison.py @@ -1,5 +1,3 @@ -#!/usr/bin/env python - # Here we will show how c_AMPs from AMPSphere are relatively similar to those in the training set of Macrel (Santos-Júnior et al., 2020) and DRAMP 3.0 (Shi et al., 2021). To that, we will use pre-computed features calculated using Macrel's internal scripts from the peptide sequences in those three databases. import pandas as pd @@ -16,7 +14,7 @@ makedirs('figures', exist_ok=True) plt.rcParams['svg.fonttype'] = 'none' plt.rcParams['font.family'] = 'sans-serif' -plt.rcParams['xtick.labelsize'] = 'smaller' +plt.rcParams['xtick.labelsize'] = 7 ampsphere = pd.read_csv('../data_folder/ampsphere_v2022-03.features.tsv.gz', sep="\t") dramp = pd.read_csv('../data_folder/dramp_v3.features.tsv.gz', sep="\t") @@ -66,10 +64,16 @@ def getlims(feat): xmax = xs[-cut] return xmin, xmax +for f in ['smallAA', 'basicAA', 'polarAA', 'aromaticAA', 'chargedAA']: + for feats in [ampsphere, dramp, macrel, neg_macrel]: + feats[f] *= 100 panels = [ ('length', 'Length (residues)'), - ('smallAA', 'Small residues'), - ('basicAA', 'Basic residues'), + ('smallAA', 'Small residues (%)'), + ('basicAA', 'Basic residues (%)'), + ('polarAA', 'Polar residues (%)'), + ('aromaticAA', 'Aromatic residues (%)'), + ('chargedAA', 'Charged residues (%)'), ('pI', 'Isoelectric point'), ('charge', 'Charge at pH 7.0'), ('aindex', 'Aliphatic index'), @@ -77,20 +81,22 @@ def getlims(feat): ('boman', 'Boman index'), ('hmoment', 'Hydrophobic\nmoment'), ('hydro_ratio','Hydrophobic ratio'), + ('hydrophobicity','Hydrophobicity'), ('flexibility','Flexibility'), ('gravy', 'GRAVY'), ] PANEL_WIDTH = 0.2 -SUBPANEL_HEIGHT = 0.07 +SUBPANEL_HEIGHT = 0.05 + fig = plt.figure() for ix,(feat,label) in enumerate(panels): pos_i = ix % 4 - pos_j = ix // 4 - ax0 = fig.add_axes([.025 + pos_i * .25, 0.075 + pos_j * .3, PANEL_WIDTH, SUBPANEL_HEIGHT]) - ax1 = fig.add_axes([.025 + pos_i * .25, 0.100 + pos_j * .3, PANEL_WIDTH, SUBPANEL_HEIGHT], sharex=ax0) - ax2 = fig.add_axes([.025 + pos_i * .25, 0.125 + pos_j * .3, PANEL_WIDTH, SUBPANEL_HEIGHT], sharex=ax0) - ax3 = fig.add_axes([.025 + pos_i * .25, 0.150 + pos_j * .3, PANEL_WIDTH, SUBPANEL_HEIGHT], sharex=ax0) + pos_j = 3 - ix // 4 + ax0 = fig.add_axes([.025 + pos_i * .25, 0.075 + pos_j * .23, PANEL_WIDTH, SUBPANEL_HEIGHT]) + ax1 = fig.add_axes([.025 + pos_i * .25, 0.100 + pos_j * .23, PANEL_WIDTH, SUBPANEL_HEIGHT], sharex=ax0) + ax2 = fig.add_axes([.025 + pos_i * .25, 0.125 + pos_j * .23, PANEL_WIDTH, SUBPANEL_HEIGHT], sharex=ax0) + ax3 = fig.add_axes([.025 + pos_i * .25, 0.150 + pos_j * .23, PANEL_WIDTH, SUBPANEL_HEIGHT], sharex=ax0) for data,data_label,ax,c in zip([neg_macrel, macrel, dramp, ampsphere], ['Macrel (neg)', 'Macrel (pos)', 'DRAMP', 'AMPSphere'], [ax0, ax1, ax2, ax3], @@ -100,6 +106,7 @@ def getlims(feat): sns.kdeplot(ax=ax, fill=True, bw_method='silverman', + bw_adjust=(2. if feat.endswith('AA') else 1.), cut=0, data=data, x=feat,