Skip to content

Results

get_tophit(result_tsv, structures, cath=False)

Process Foldseek output to extract top hit and weighted bitscores.

Parameters:

Name Type Description Default
result_tsv Path

Path to the Foldseek result TSV file.

 required
structures bool

Flag indicating whether structures have been added.

 required
cath bool

Flag indicating whether this is for CATH database (all greedy besthits kept not just top)

 False

Returns:

Type Description
Tuple[pd.DataFrame, pd.DataFrame]

Tuple[pd.DataFrame, pd.DataFrame]: A tuple containing two DataFrames: 1. DataFrame containing the top functions extracted from the Foldseek output. 2. DataFrame containing weighted bitscores for different functions.
Source code in src/baktfold/results/tophit.py 
 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
def get_tophit(
    result_tsv: Path,
    structures: bool,
    cath: bool = False
) -> Tuple[pd.DataFrame, pd.DataFrame]:
    """
    Process Foldseek output to extract top hit and weighted bitscores.

    Args:
        result_tsv (Path): Path to the Foldseek result TSV file.
        structures (bool): Flag indicating whether structures have been added.
        cath (bool): Flag indicating whether this is for CATH database (all greedy besthits kept not just top)

    Returns:
        Tuple[pd.DataFrame, pd.DataFrame]: A tuple containing two DataFrames:
            1. DataFrame containing the top functions extracted from the Foldseek output.
            2. DataFrame containing weighted bitscores for different functions.
    """

    logger.info("Processing Foldseek output")

    if structures:

        col_list = [
            "query",
            "target",
            "bitscore",
            "fident",
            "evalue",
            "qStart",
            "qEnd",
            "qLen",
            "tStart",
            "tEnd",
            "tLen",
            "alntmscore",
            "lddt"
        ]
    else:

        col_list = [
            "query",
            "target",
            "bitscore",
            "fident",
            "evalue",
            "qStart",
            "qEnd",
            "qLen",
            "tStart",
            "tEnd",
            "tLen",
        ]

    foldseek_df = pd.read_csv(
        result_tsv, delimiter="\t", index_col=False, names=col_list
    )

    # replace ~PIPE~ with |
    foldseek_df["query"] = foldseek_df["query"].str.replace("~PIPE~", "|", regex=False)


    # in case the foldseek output is empty
    if foldseek_df.empty:
        logger.warning(
            "Foldseek found no hits whatsoever - please check your input if you expect hits"
        )

    else:

        # add qcov and tcov 
        foldseek_df["qCov"] = ((foldseek_df["qEnd"] - foldseek_df["qStart"] ) / foldseek_df["qLen"]).round(2)
        foldseek_df["tCov"] = ((foldseek_df["tEnd"] - foldseek_df["tStart"] ) / foldseek_df["tLen"]).round(2)

        # reorder
        qLen_index = foldseek_df.columns.get_loc("qLen")
        tLen_index = foldseek_df.columns.get_loc("tLen")

        new_column_order = (
            list(
                [
                    col
                    for col in foldseek_df.columns[: qLen_index + 1]
                    if col not in ["qCov", "tStart","tEnd",	"tLen", "tCov"]
                ]
            )
            + ["qCov", "tStart","tEnd",	"tLen", "tCov"]
            + list(
                [
                    col
                    for col in foldseek_df.columns[tLen_index + 1 :]
                    if col not in ["qCov", "tStart","tEnd",	"tLen", "tCov"]
                ]
            )
        )
        foldseek_df = foldseek_df.reindex(columns=new_column_order)


        if not cath:
            # get only the tophit - will always be the first hit for each query (top bitscore)
            foldseek_df = foldseek_df.drop_duplicates(subset="query", keep="first")
        # otherwise, the df will contain all greedy tophits from CATH


    return foldseek_df