Hi,

due to the redundancy of the genetic code you cannot go from amino acid to codons. We developed a tool called MACSE dedicated to such tasks:

concerning alignment

If you have a nice amino acid alignment (manually corrected) you thus need to have the unaligned nucleotide sequences of the CDS corresponding to the amino acid aligned sequences to transpose your amino acid alignment at the codon level, this can be done using several tools. For instance using MACSE the command line for this will be:

java -jar macse_v1.01b.jar -prog reportGapsAA2NT -align nice_AA_alignment.fasta -seq unaligned_CDS.fasta -out_NT nice_corresponding_NT.fasta

If your amino acid alignments have been done automatically, you can use tools sur as revTrans, pal2nal or MACSE to automatized a 3 step alignment approach:

- translate your nucleotide to AA

- align those AA sequences

- report gap of this alignment to the nucleotide sequence

If some of your sequences may contain biological frameshifts, or artefactual ones (i.e. an erroneous addition/missing nucleotide, as could append in NGS contigs) the translated amino acid sequences will be erroneous after this frameshit and the whole strategy will fail. We precisely developed MACSE to provide a tool that can automatically overcome such problem by integrating the possibility of frameshift during the alignment process. To run MACSE with default options just use the following command line:

java -jar macse_v1.01b.jar -prog alignSequences -seq unaligned_CDS.fasta -out_NT aligned_CDS_NT.fasta -out_AA aligned_CDS_AA.fasta

You can obviously choose the adapted genetic code if your are dealing with ribosomal genes and tune other options for advance usage.

concerning positive selection

Your alignment may contain gaps when searching for positive selection with tools such as paml  or mapNH (http://biopp.univ-montp2.fr/forge/testnh). 

But you need to use a carefully curated alignment (either manually or using automatic tools such as trimAl, guidance, zorro or BMGE), or a manual check of identified potential positive selection locus/branches. Indeed any misaligned sequence fragment tends to lead to a positive selection signature, see for instance our recent paper that underline the impact of this artefact (http://www.biomedcentral.com/1471-2229/14/151).

I hope you will find this comment useful

http://bioweb.supagro.inra.fr/macse/

http://abacus.gene.ucl.ac.uk/software/paml.html)

I suggess this book to do many basic and advanced bioinformatic analyses "bioinfomatic for dummies" you can find the response of your question also, i remember that i red the method through my reading the book 

You can read more about the role of multiple sequence alignments (MSA) programs in biological research. Genes, genomes and genomics.

Article Assessment of molecular (dis)similarity: The role of multipl...

Use Mega 5.2 software. it may help u.

USE EXPASY PORTAL AS THIS WILL HELP YOU TRANSLATE THE PROTEINS BACK TO THEIR NUCLEOTIDES OR TRY TO INPUT THE ACCESION NUMBER OF THE PROTEIN SEQUENCE INTO THE NCBI DATA BASE AND CODE BACK TO NUCLEOTIDES AS THIS WILL AVOID THE ISSUES OF CODON USAGE BAISE. FOR OFRS USE ORF FINDER IN THE NCBI THIS WILL GIVE SATISFACTORY RESULTS.

BEST WISHES SIGNH.

You can use PAL2NAL tool (http://www.bork.embl.de/pal2nal/index.cgi#RunP2N) to convert protein alignment into nucleotide alignment. You need to give protein alignment file and cds/cdna sequences for same set of proteins in same order. the program will give you nucleotide alignment as an output in the format you have chosen in the settings of the program.  

There is also the RevTrans server (http://www.cbs.dtu.dk/services/RevTrans-2.0/web/) which takes coding DNA sequences, translates them to amino acids and aligns them at the protein level using a choice of alignment methods, then maps the DNA sequences to the alignment. It can also use existing protein alignments.

http://www.cbs.dtu.dk/services/RevTrans-2.0/web/

Hi,

due to the redundancy of the genetic code you cannot go from amino acid to codons. We developed a tool called MACSE dedicated to such tasks:

concerning alignment

If you have a nice amino acid alignment (manually corrected) you thus need to have the unaligned nucleotide sequences of the CDS corresponding to the amino acid aligned sequences to transpose your amino acid alignment at the codon level, this can be done using several tools. For instance using MACSE the command line for this will be:

java -jar macse_v1.01b.jar -prog reportGapsAA2NT -align nice_AA_alignment.fasta -seq unaligned_CDS.fasta -out_NT nice_corresponding_NT.fasta

If your amino acid alignments have been done automatically, you can use tools sur as revTrans, pal2nal or MACSE to automatized a 3 step alignment approach:

- translate your nucleotide to AA

- align those AA sequences

- report gap of this alignment to the nucleotide sequence

If some of your sequences may contain biological frameshifts, or artefactual ones (i.e. an erroneous addition/missing nucleotide, as could append in NGS contigs) the translated amino acid sequences will be erroneous after this frameshit and the whole strategy will fail. We precisely developed MACSE to provide a tool that can automatically overcome such problem by integrating the possibility of frameshift during the alignment process. To run MACSE with default options just use the following command line:

java -jar macse_v1.01b.jar -prog alignSequences -seq unaligned_CDS.fasta -out_NT aligned_CDS_NT.fasta -out_AA aligned_CDS_AA.fasta

You can obviously choose the adapted genetic code if your are dealing with ribosomal genes and tune other options for advance usage.

concerning positive selection

Your alignment may contain gaps when searching for positive selection with tools such as paml  or mapNH (http://biopp.univ-montp2.fr/forge/testnh). 

But you need to use a carefully curated alignment (either manually or using automatic tools such as trimAl, guidance, zorro or BMGE), or a manual check of identified potential positive selection locus/branches. Indeed any misaligned sequence fragment tends to lead to a positive selection signature, see for instance our recent paper that underline the impact of this artefact (http://www.biomedcentral.com/1471-2229/14/151).

I hope you will find this comment useful

http://bioweb.supagro.inra.fr/macse/

http://abacus.gene.ucl.ac.uk/software/paml.html)

also vote for pal2nal

use pal2nal 

If you have many alignments to check, go for local setup rather than using the server tool. That will give you batch conversion advantage.

Check the attached paper

Article Identification of the positively selected genes governing ho...

pal2nal is good--you might also use Gblocks to remove iffy regions from your alignment.

http://molevol.cmima.csic.es/castresana/Gblocks_server.html

I agree with Vincent : "due to the redundancy of the genetic code you cannot go from amino acid to codons" ... but, there are several ways to guess the "most probable" codon set, even if you don't have the DNA sequence or don't want to go back to ...

We have developed a tool named PATH (see the link below) for pairwise protein alignments with frameshift and without genomic/DNA sequence that may help. We have found interesting conserved frame-shifts in protein sequences where double encoding existed "for a while" ... and this not only in Bacterial or Viruses genomes :-)

Have a nice day

L.

http://bioinfo.lifl.fr/path

Article Back-translation for discovering distant protein homologies ...