如何将字符串拆分为字符并用浮点值替换字符以找到 Python 中原始字符串的总和？

Question

Hi,

Noobie to python here.

I have >10,000 strings that represent peptide sequences. Each letter in the string is an amino acid and I would like to calculate the "net sum" of the string after I have replaced each letter with a pre-defined float value (ranging from -1 to -2).

I am stuck on where to start with the loop to make this work? I have the code to clean the strings so that non-alphabetical characters are removed and replace with float values defined in a dictionary (ie W:2.10, G:-1.0)

cleaned peptides, truncated to 5 characters

I imagine the code is something like.

I have 6 dataframes to repeat this process in.

Any help would be immensely appreciated!

Updated Code (THIS WORKS THANKS TO SARAH MESSER)

def hydrophobicity_score(peptide):
    hydro = { 
        'A': -0.5,
        'C': -1.0,
        'D': 3.0,
        'E': 3.0,
        'F': -2.5,
        'G': 0.0,
        'H': -0.5,
        'I': -1.8,
        'K': 3.0,
        'L': -1.8,
        'M': -1.3,
        'N': 0.2,
        'P': 0.0,
        'Q': 0.2,
        'R': 3.0,
        'S': 0.3,
        'T': -0.4,
        'V': -1.5,
        'W': -3.4,
        'Y': -2.3,
    }
    hydro_score = [hydro.get(aa,0.0)for aa in peptide]
    return sum(hydro_score)

og_pep['Hydro'] = og_pep['Peptide'].apply(hydrophobicity_score)
og_pep

Answer 1

Okay, first up, you don't want to loop over the rows in a dataframe. The rows are designed to be processed in parallel. Getting your head around that is a bit of a stretch, but once you've defined a few row-level operations and applied them to large dataframes, it'll get smoother. (The problem with looping over rows is one of speed . It's sometimes useful in debugging or toy problems, but modern computing hardware tries to parallelize computations as much as possible. Dataframes take advantage of that to process all the rows at once, rather than handling them individually in a loop.)

To do the conversion, you're going to need to define a custom function to operate on each individual row. Then you pass that custom function to the dataframe and tell it to apply that row-level function to one column in order to generate a new column.

So here's a possible function to get you started:

def peptide_score(peptide_string):
    '''Returns a numerical score given a sequence of peptide characters.'''
    # Replace the values in this dict (dictionary / map) with whatever values you need
    amino_acid_scores = { 
        'A': 0.1,
        'C': 1.4,
        'G': 0.32342,
        'T': -0.23,
        'U': 74.22
    }
    # This is called a "list comprehension." It's great for transforming sequences.
    score_list = [amino_acid_scores[character] for character in peptide_string]
    return sum(score_list)

# I'm assuming your pre-existing dataframe is called "gluc_dataframe" and that the
# column with your strings is called "Peptide".  Output scores will be stored in a new
# column, "score". Replace those names with whatever fits.
gluc_dataframe['score'] = gluc_dataframe['Peptide'].apply(peptide_score)

If you've got a lot of characters you want to ignore (whitespace, punctuation, whatever), you can replace amino_acid_scores[character] in the list comprehension with amino_acid_scores.get(character, 0.0) .

Answer 2

def hydrophobicity_score(peptide):
     hydro = { 
        'A': -0.5,
        'C': -1.0,
        'D': 3.0,
        'E': 3.0,
        'F': -2.5,
        'G': 0.0,
        'H': -0.5,
        'I': -1.8,
        'K': 3.0,
        'L': -1.8,
        'M': -1.3,
        'N': 0.2,
        'P': 0.0,
        'Q': 0.2,
        'R': 3.0,
        'S': 0.3,
        'T': -0.4,
        'V': -1.5,
        'W': -3.4,
        'Y': -2.3,
    }
    hydro_score = [hydro[aa] for aa in peptide]
    return sum(hydro_score)

og_peptide= og_pep['Peptide']
og_peptide = og_peptide.str.replace('\W+','')
og_peptide = og_peptide.str.replace('\d+','')
og_peptide = pd.DataFrame(og_peptide)
og_peptide['Hydro_Score'] = og_peptide.apply(hydrophobicity_score)
og_peptide

I am not getting the expected output.

Output

Here is og_pep DataFrame