ImageJ (Fiji) STORM microscopy autocorrelation analysis

autocorrel_analysis.py

import pandas as pd
import os
import sys

csv_autocor_name = "Recs TS proc (xy16).tif_AC(l1,w0)_Autocorr.csv"


def export_to_csv(sheets, analyzis_path):
    for k, v in sheets.items():
        v.to_csv(os.path.join(analyzis_path, k + ".csv"), index=None)


def export_to_excel(sheets, doc_name):
    with pd.ExcelWriter(doc_name) as writer:
        for k, v in sheets.items():
            v.to_excel(writer, sheet_name=k, index=None)


def import_data(csv_path):
    df_csv = pd.read_csv(os.path.abspath(csv_path))
    first_col = df_csv.iloc[:, 0]
    headers = df_csv.columns.to_list()[1:]
    cond_names = [h.split(sep="_")[0] + " " + h.split(sep="_")[2] for h in headers]
    conds_indexes = dict(enumerate([c for c in cond_names]))
    conds_prefixes = sorted(set(conds_indexes.values()))
    sheets = dict((cp, pd.DataFrame()) for cp in conds_prefixes)  # create one df per sheet (or condition)

    # put data of each condition in a distinct dataframe
    roi_idx = 1
    for idx, cond in conds_indexes.items():
        col = df_csv.iloc[:, idx+1]
        header_mod = str.join(" ", ["ROI" + str(roi_idx).zfill(3), str.join(" ", col.name.split(sep="_")[:3])])
        sheets[cond][first_col.name] = first_col.tolist()
        sheets[cond][header_mod] = col.tolist()
        roi_idx += 1

    return sheets


def process_data(sheets, conditions=[]):
    df = pd.DataFrame()
    df_m = pd.DataFrame()

    for idx, (k, v) in enumerate(sheets.items()):
        col_name = conditions[idx] if conditions and len(conditions) == len(sheets) else k
        valleys1 = v.iloc[6, 1:].values  # line 6 is 0.096 µm
        valleys2 = v.iloc[18, 1:].values  # line 18 is 0.288 µm
        valleys_mean = (valleys1 + valleys2) / 2
        peaks = v.iloc[12, 1:].values  # line 12 is 0.192 µm
        df_p2v = pd.DataFrame({col_name: peaks - valleys1})
        df_p2v_m = pd.DataFrame({col_name + "_m": peaks - valleys_mean})
        df = pd.concat([df, df_p2v], axis=1)
        df_m = pd.concat([df_m, df_p2v_m], axis=1)

    return df, df_m


def format_data(sheets, result):
    sheets["peak-valley"] = result[0]
    sheets["peak-valleys_mean"] = result[1]
    return sheets


if __name__ == '__main__':
    if len(sys.argv) > 2:
        xp_path = sys.argv[1]
        AIS_subdir = sys.argv[2]
        analyzis_path = os.path.join(xp_path, AIS_subdir)
        ac_csv_path = os.path.join(analyzis_path, csv_autocor_name)

        sheets = import_data(ac_csv_path)
        result = process_data(sheets)
        sheets = format_data(sheets, result)
        export_to_csv(sheets, analyzis_path)

    else:
        print("not enough args", sys.stderr)
        print("usage: python autocorrel_analysis.py /path/to/experiment_dir subfolder")
        print("where subfolder is in experiment_dir and contains the CSV resulting from ImageJ autocorrel analysis")
        print("this software assumes file name formatting with prefix like 'Cx_Ny_condname_...'")