how to align lists of words using biopython pairwise2

Question

[euacar]

When I run the script below, output is getting split into single chars. Any idea why? It looks like the second argument gets split into single chars. I am trying to align the word sequences. I will have many words hence cannot map them to letters only.

Thx Ercan

from Bio.Seq import Seq

from Bio.pairwise2 import format_alignment

fruits = ["orange","pear", "apple","pear","orange"]

fruits1 = ["pear","apple"]


from Bio import pairwise2

alignments = pairwise2.align.localms(fruits,fruits1,2,-1,-0.5,-0.1, gap_char=["-"])

for a in alignments: 

    print(format_alignment(*a))

Output:

 ['orange', 'r', 'a', 'e', 'p', 'e', 'l', 'p', 'p', 'a', 'pear', 'orange']
 |||||||||
['-', 'r', 'a', 'e', 'p', 'e', 'l', 'p', 'p', 'a', '-', '-']
  Score=4

[Joe]

Please reformat your code using the 101010 button or by putting 4 spaces before each line of code. This is particularly important for python as you'll need to correctly indent the code so that we can be sure it's written correctly.

[euacar]

Thx For the feedback

[WouterDeCoster]

As said by jrj.healey, code formatting is very important. I now added code markup to your post for increased readability. You can do this by selecting the text and clicking the 101010 button. When you compose or edit a post that button is in your toolbar, see image below:

[euacar]

Thx For the feedback

Answer 1

[Andrzej Zielezinski]

BioPython allows for performing an alignment on sequences' symbols (not words). I can't think of any program that would do the words' alignment. I suggest you search the Web for Python modules/functions that perform string character alignments and then modify the existing code to handle words.

For example, I tweaked some code from GitHub to do local (water) and global (needle) alignments of words.

Filename: alignment.py:

Usage:

import alignment
fruits1 = ["orange", "pear", "apple", "pear", "orange"]
fruits2 = ["pear", "apple"]
aln = alignment.needle(fruits1, fruits2)
identity = aln[0]
score = aln[1]
print(identity, score)
print(aln[2])
print(aln[3])

Output:

(0.4, 25)
['-', 'pear', 'apple', '-', '-']
['orange', 'pear', 'apple', 'pear', 'orange']

[euacar]

Thank you for the code..

Answer 2

[Markus]

You can do this with Biopython's pairwise2 (again). There was a bug in pairwise2 in Biopython releases 1.68/1.69 which prevented the proper handling of lists as input. If you run the same code in Biopython 1.70 you get the following (and expected) result:

['orange', 'pear', 'apple', 'pear', 'orange']
 ||
['-', 'pear', 'apple', '-', '-']
  Score=4

[Andrzej Zielezinski]

Thanks for posting this info. Good to know that.

Answer 3

[yannis1962]

Comparing two lists of words didn't worked for me either, so what I did was to convert words into Chinese characters (there are more than 20,000 of them in Unicode), aligning the sequences as character strings, and then back to Latin-alphabet words again. Works like a charm:

from bio import pairwise2
from bio.pairwise2 import format_alignment

LISTA=["alors","en","fait","depuis","novembre","2017","du","coup",",","j'","ai","fait",",","plusieurs","ce","que","j'","appelle","des","crises",",","en","fait","c'","est",",","pendant",",","une","semaine",",","entre","une","semaine","et","10","jours",",","je","me","sens","un","peu","comme",",","déconnectée","de","la","réalité","un","peu",",","j'","arrive","pas","à",",","j'","arrivais","plus","à","faire","la","différence","entre",",","entre","si","j'","étais","dans","un","ou","si","j'","étais","dans","la","réalité","."]
LISTB=["alors","en","fait","depuis","euh","novembre","deux","mille","dix-sept","du","coup","j'","ai","fait","euh","hum","hum","plusieurs","euh","enfin","ce","que","j'","appelle","des","crises","en","fait","c'","est","euh","pendant","euh","une","semaine","entre","une","semaine","et","dix","jours","euh","je","me","sens","un","peu","comme","euh","déconnectée","de","la","réalité","un","peu","j'","arrive","pas","à","j'","arrivais","plus","à","faire","la","différence","entre","entre","si","j'","étais","dans","un","rêve","ou","si","j'","étais","dans","la","réalité"]

charcode=ord(u"一")-1
LATtoHAN={}
HANtoLAT={}
LISTA_=[]
LISTB_=[]
for x in LISTA:
    if x in LATtoHAN.keys():
        LISTA_.append(LATtoHAN[x])
    else:
        charcode+=1
        LATtoHAN[x]=chr(charcode)
        HANtoLAT[chr(charcode)]=x
        LISTA_.append(LATtoHAN[x])
for x in LISTB:
    if x in LATtoHAN.keys():
        LISTB_.append(LATtoHAN[x])
    else:
        charcode+=1
        LATtoHAN[x]=chr(charcode)
        HANtoLAT[chr(charcode)]=x
        LISTB_.append(LATtoHAN[x])

LISTA__="".join(LISTA_)
LISTB__="".join(LISTB_)

alignments=pairwise2.align.globalxx(LISTA__,LISTB__)

RESA=[]
RESB=[]
for w in alignments[0][0]:
    if (w == "-"):
        RESA.append("-")
    else:
        RESA.append(HANtoLAT[w])
for w in alignments[0][1]:
    if (w == "-"):
        RESB.append("-")
    else:
        RESB.append(HANtoLAT[w])

print(RESA,RESB)