I have had to do something very similar recently.

The approach to solving your problem requires 3 steps. Here is my take:

• Use Python with Biopython to pass through your sequence couples, one individual at a time. Reverse complement the second sequence. (use any other preferred approach)
• Use the "dialign2" program repeatedly for each individual to create alignments. In this way, your 2 sequences will have the same length (with "-" characters to complete the missing parts).
• Use Python to pass through each alignment (2 sequences each time) and create a consensus. This should be easy.

Here is an attempt for the third part:

def create_consensus(seq1, seq2):
    """Sequences must be strings, have the same length, and be aligned"""
    out_seq = ""
    for i, nucleotide in enumerate(seq1):
        couple = [nucleotide, seq2[i]]
        if couple[0] == "-":
            out_seq += couple[1]
        elif couple[1] == "-":
            out_seq += couple[0]
        elif couple[0] == couple[1]:
            out_seq += couple[0]
        elif not couple[0] == couple[1]:
            out_seq += "N"
    return out_seq

seq1 = "-----CAGATACGCGCACATAGACATAG"
seq2 = "ACTGACAGATACAGACACATA-------"

print seq1
print seq2
print create_consensus(seq1, seq2)

The printed result is:

-----CAGATACGCGCACATAGACATAG
ACTGACAGATACAGACACATA-------
ACTGACAGATACNNNCACATAGACATAG

This is admittedly quickly done and there may be both 1) mistakes or special cases 2) a better way to do this, maybe with Biopython.

Hope it helps!

Cheers

Following up on Eric's useful answer, there is functionality to do this in Biopython. The Tutorial provides details on using it: http://biopython.org/DIST/docs/tutorial/Tutorial.html

• Chapter 6 deals with reading in alignment objects
• Chapter 14.3 provides useful functionality once you have an alignment object.

Specifically the function "dumb_consensus" calculates a consensus along the lines of what you are doing with create_consensus, with a bit of extra functionality like defining when you call something ambiguous. Source code for that is here:

http://github.com/biopython/biopython/blob/master/Bio/Align/AlignInfo.py

I worked on a solution to this that is available here - https://raw.github.com/gawbul/bioinformatics-scripts/master/FRconsensus.py

Cheers,
Steve

Pagan has a consensus option that does what you need: http://www.ebi.ac.uk/~ari/pagan/