MarkerIntensityPredictor/predictive_performance_per_round.py at main · goeckslab/MarkerIntensityPredictor · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
import argparse
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
from pathlib import Path

rounds = {"9_2": {
    "2": ["pERK"],
    "3": ["Vimentin", "aSMA"],
    "4": ["Ecad", "ER", "PR"],
    "5": ["EGFR", "pRB", "CD45"],
    "6": ["Ki67", "CK19", "p21"],
    "7": ["CK14", "AR"],
    "8": ["CK17", "HER2"],
},
    "9_3": {
        "2": ["pERK"],
        "3": ["Vimentin", "aSMA"],
        "4": ["Ecad", "ER", "PR"],
        "5": ["EGFR", "pRB", "CD45"],
        "6": ["Ki67", "CK19", "p21"],
        "7": ["CK14", "AR"],
        "8": ["CK17", "HER2"],
    },
    "9_14": {
        "2": ["pERK"],
        "3": ["Ki67", "CD45"],
        "4": ["Ecad", "aSMA", "Vimentin"],
        "5": ["pRB", "EGFR", "p21"],
        "7": ["ER", "HER2"],
        "8": ["CK14", "CK19", "CK17"],
        "9": ["AR"],
        "10": ["PR"]
    },
    "9_15": {
        "2": ["pERK"],
        "3": ["Ki67", "CD45"],
        "4": ["Ecad", "aSMA", "Vimentin"],
        "5": ["pRB", "EGFR", "p21"],
        "7": ["ER", "HER2"],
        "8": ["CK14", "CK19", "CK17"],
        "9": ["AR"],
        "10": ["PR"]
    }
}

results_folder = Path("results/scores_by_round")

if __name__ == '__main__':
    if not results_folder.exists():
        results_folder.mkdir(parents=True)

    # Check whether marker expression is high

    parser = argparse.ArgumentParser()
    parser.add_argument('-b', "--biopsy", help='Input file')
    parser.add_argument("-hyper", "--hyper", action="store_true", default=False)
    parser.add_argument("-f", "--fe", action="store_true", default=False)
    parser.add_argument("--snr", action="store_true", default=True)
    args = parser.parse_args()

    biopsy = args.biopsy
    segmentation = biopsy.split("/")[1]
    print(segmentation)
    if segmentation == "tumor_mesmer":
        segmentation = "Mesmer"
    elif segmentation == "tumor_s3_snr":
        segmentation = "Unmicst_+_S3"

    hyper = 0 if not args.hyper else 1
    fe = None if not args.fe else "SP"

    # Load score data from data scores folder

    scores = []
    for model in ["EN", "Ludwig"]:
        snr = 0 if not args.snr else 1
        scores.append(pd.read_csv(Path("data/scores",
                                       f"{Path(args.biopsy).stem}_IP_{'_'.join(segmentation.split(' '))}_{snr}_{fe}_{model}_{hyper}_scores.csv")))
        scores.append(pd.read_csv(Path("data/scores",
                                       f"{Path(args.biopsy).stem}_OP_{'_'.join(segmentation.split(' '))}_{snr}_{fe}_{model}_{hyper}_scores.csv")))

    assert len(scores) == 4, "Not all scores are loaded"
    scores = pd.concat(scores)
    # create new column which merges type and model together
    scores["Type"] = scores["Type"] + "_" + scores["Mode"]
    # replace _ with spaces for Type column
    scores["Type"] = scores["Type"].str.replace("_", " ")
    case = "_".join(Path(biopsy).stem.split("_")[0:2])
    rounds = rounds[case]

    # plot boxen plot for each round

    # plot boxen plot for df, with hue round
    fig, axes = plt.subplots(nrows=2, ncols=4, figsize=(10, 5), dpi=200)
    row = 0
    col = 0

    score = "MAE"
    for round in rounds:
        ax = axes[row][col]
        markers = rounds[round]
        # Select all markers for round and mae scores
        df = scores[scores["Marker"].isin(markers)][["Marker", score, "Type"]]

        sns.barplot(data=df, x="Marker", y="MAE", hue="Type", ax=ax)

        ax.legend_ = None
        # set title of ax
        ax.set_title(f"Round {round}")
        # set x label of ax
        ax.set_xlabel("Markers")
        # set y label of ax
        ax.set_ylabel(score)
        # set ylim of ax
        ax.set_ylim(0, 0.55)

        if col == 3:
            row += 1
            col = 0
        else:
            col += 1

    if len(rounds) != 8:
        fig.delaxes(axes[row, col])

    # set figure wide title
    fig.suptitle(
        f"Shared Marker MAE per round\n{segmentation.replace('_', ' ')} {'SNR' if snr else ''} \nBiopsy {Path(biopsy).stem.replace('_', ' ')}")

    # set bboxes of legend outside of figure
    # use axes legend attributes to create a fig legen
    fig.legend(*ax.get_legend_handles_labels(), loc="lower right", bbox_to_anchor=(1.001, 0.2))
    plt.tight_layout()

    plt.savefig(
        Path(results_folder, f"{Path(biopsy).stem}_{segmentation}_{snr}_marker_prediction_per_round.png"))