Hello.
I have two genes with a homology of 96%.
I want to align my PCR product after doing next-generation sequencing to human reference genome to a specific gene (one of the two) but not to the other.
Is there any solution to avoid an alignment for the unwanted gene? Masking or something similar?
Thanks a lot.
First, homology of XX%, what is that? please check this paper: http://www.ncbi.nlm.nih.gov/pubmed/10782117
What you should be referring to is 96% sequence identity.
It would be impossible to assign each read to its originating gene. Most reads will align equally well to either of the two genes, only reads where there are polymorphic sites will help you here. I am assuming here that your PCR primers will amplify any of the two genes. If that is not the case, and your primers only amplify the gene of interest, then you can mask the location of the other gene in the genome sequence, i.e., replace the sequence by Ns. You could use BEDTools's maskFastaFromBed to mask the non-desired region, check: http://bedtools.readthedocs.org/en/latest/content/tools/maskfasta.html
Hope this helps.
Well, in a crude approach, you could use Clustal Omega from the EBI website (http://www.ebi.ac.uk/Tools/msa/clustalo/).
Just prepare a text file containing the sequence of the only gene you are interested in in fasta format (so with the ">" sign in first line) and the sequencing data in same format (after the ">" sign you can write an id. name/number for the sequence and it will appear in the alignment) and then upload it in the webpage, select DNA to specify the sequence type, and run the alignment.
With short NGS reads and high similarity between your two genes, it is very likely that reads from gene A will align to the reference sequence from gene B, even when the reads do not match for 100%. In this case, you will get false positive heterozygous variations (which are not real variations of your gene A or B but is a paralogous sequence variant).
Segmental duplications and highly similar genes are tricky. Carefully examine your reads (for example in IGV).
Good luck!
You may look at this publication:
http://genomebiology.com/2011/12/6/R55
With that sequence homology (>95%), I have doubt bioinformatics can handle this problem. You may have to do Sanger sequencing (choosing correct primers) if the exons of pseudogenes (paralogs) are processed. Otherwise you would like to do long-range PCR and subsequent nested PCR for sequencing - if the sequences for both the genes are >95% similar in exons/introns/ and Ex-Int boundary regions.
First, homology of XX%, what is that? please check this paper: http://www.ncbi.nlm.nih.gov/pubmed/10782117
What you should be referring to is 96% sequence identity.
It would be impossible to assign each read to its originating gene. Most reads will align equally well to either of the two genes, only reads where there are polymorphic sites will help you here. I am assuming here that your PCR primers will amplify any of the two genes. If that is not the case, and your primers only amplify the gene of interest, then you can mask the location of the other gene in the genome sequence, i.e., replace the sequence by Ns. You could use BEDTools's maskFastaFromBed to mask the non-desired region, check: http://bedtools.readthedocs.org/en/latest/content/tools/maskfasta.html
Hope this helps.
- Badges
- Science method