如何有条件地替换子列表中的项目。

Question

I am attempting to generate a list-within-a-list. I am cycling through a file to update the list if one sublist element is greater. I have written this code:

targets = open(file)

longest_UTR = []

for line in targets:

    chromosome, locus, mir, gene, transcript, UTR_length = line.strip("\n").split("\t")

    length_as_integer = int(UTR_length)

    if not any(x[:3] == [locus, mir, gene] for x in longest_UTR):

        longest_UTR.append([locus, mir, gene, transcript, length_as_integer])

    elif length_as_integer > [int(x[4]) for x in longest_UTR]: ##x[4] = previous length_as_integer

        longest_UTR.append([locus, mir, gene, transcript, length_as_integer])

print (longest_UTR)

However, I get this error:

elif len_as_int > (int(x[4]) for x in longest_UTR):

TypeError: '>' not supported between instances of 'int' and 'generator'

How can I convert x[4] to an integer so as to compare to length_as_integer ?

Thank you

Answer 1

If I get this right, try replacing the elif line with the following:

else:
    longest_UTR = [[locus, mir, gene, transcript, length_as_integer] for x in longest_UTR if x[:3] == [locus, mir, gene] and length_as_integer > int(x[4]) else x]:

You pass through all your list updating the ones matching the condition and doing nothing if it doesn't match.

Answer 2

So, there's been a bit of back and forth regarding your requirements, but my final understanding is this: You are looping over a data set. Each target in this data set has a locus , mri , and gene as well as a UTR_length attribute. For every unique combination of locus , mri , and gene you are trying to find all targets that have the maximum UTR_Length ?

Given that you are wanting to find the maximum value in the dataset there are two approaches.
1) You could simply convert your input file to a pandas dataframe, group by you locus , mri and gene values, and return all values with max( UTR_Length ). From ease of implementation this is probably your best bet. However, pandas is not always the right tool, and carries a lot of overhead, especially if you want to Dockerise your project.

2) If you want to use base python packages, I would recommend taking advantage of sets and dictionaries:

targets = open(file)
list_of_targets = []    
for line in targets:

          chromosome, locus, mir, gene, transcript, UTR_length = line.strip("\n").split("\t")
          length_as_integer = int(UTR_length)

          list_of_targets.append((chromosome, locus, mir, gene, transcript, UTR_length))

# Generate Set of unqiue locus, mri, gene (lmg) combinations
set_of_locus_mri_gene = {(i[1], i[2], i[3]) for i in list_of_targets}

# Generate dictionary of maximum lengths for each distinct lmg combo
dict_of_max_lengths = {lmg: max([targets[5] for targets in list_of_targets if 
                                    (targets[1], targets[2], targets[3]) == lmg]) for 
                                    lmg in set_of_locus_mri_gene}

# Generate dictionary with lmg keys and all targets with corresponding max length
final_output = {lmg: [target for target in list_of_targets if target[5] == max_length] for
                        lmg, max_length in dict_of_max_lengths.items()}

Answer 3

Since you want to replace the longest_UTR variable and keep things nicely named you could use a dictionary instead of a list:

targets = open(file)
longest_UTR = {}

for line in targets: 
    chromosome, locus, mir, gene, transcript, UTR_length = line.strip("\n").split("\t")    
    length_as_integer = int(UTR_length)

    # Your condition works for initializing the dictionary because of the default value.
    if length_as_integer > longest_UTR.get("Length", -1):
        longest_UTR["Chromosome"] = chromosome
        longest_UTR["Locus"] = locus
        longest_UTR["Mir"] = mir
        longest_UTR["Gene"] = gene
        longest_UTR["Transcript"] = transcript
        longest_UTR["Length"] = length_as_integer

print (longest_UTR)

Edit: here is also the version of the code using a list, just in case you are interested to see the difference. Personally I find the dictionary approch cleaner to read.

targets = open(file)
longest_UTR = [None, None, None, None, None, -1]

for line in targets: 
    chromosome, locus, mir, gene, transcript, UTR_length = line.strip("\n").split("\t")    
    length_as_integer = int(UTR_length)

    # Your condition works for initializing the list because of the default value.
    if length_as_integer > longest_UTR[5]:
        longest_UTR[0] = chromosome
        longest_UTR[1] = locus
        longest_UTR[2] = mir
        longest_UTR[3] = gene
        longest_UTR[4] = transcript
        longest_UTR[5] = length_as_integer

print (longest_UTR)