3. Aligning Sequences

Generating an Automatic Alignment

Now we'll generate an automatic alignment we can start from in building our RNA family. Feed the fasta file you put together in the last section to WAR. Alternatively you can use the example fasta file below. Precomputed results for this set are available here. Note: if WAR is down, the EMBL alignment servers are an alternative here.

>U00096.2

GAAAGACGCGCATTTGTTATCATCATCCCTGAATTCAGAGATGAAATTTTGGCCACTCACGAGTGGCCTTTTT

>FQ312003

GAAAGACGCGCATTTGTTATCATCATCCCTGTTTTCAGCGATGAAATTTTGGCCACTCCGTGAGTGGCCTTTTT

>CP002272

GAAAGACGCGCATTTGTTATCATCATCCCTGACTTCAGAGATGAAATGTTTGGCCACAGTGATGTGGCCTTTTT

>CP002910

GAAAGACGCGCATTTATTATCATCATCATCCCTGAATCAGAGATGAAAGTTTGGCCACAGTGATGTGGCCTTTTT

>AM286415

GAAAGACGCGCATTTGTTATCATCATCCCTGTTATCAGAGATGTTAATTTGGCCACAGCAATGTGGCCTTTT

>CP002433

GAAAGACGCGCATTTGTTATCATCATCCCTGACAACAGAGATGTTAATTCGGCCACAGTGATGTGGCCTTTT

>FP236842

GAAAGACGCGTATTTGTTATCATCATCTCATCCCTGACAACAGAGATGTTAATTTAGGCCACAGTGACGTGGCCTTTTT

Once you have your results, take a look at the different alignments and secondary structures predicted. In particular the T-coffee consensus combines results from all the other methods, and may give you a good idea of regions where manual refinement my improve the alignment. Usually you'll want to use the consensus alignment, but other alignments may be better if a majority appear to predict "odd" secondary structures.

Manual Refinement

Open the alignment use RALEE. First look at the structure. Are there base-pairs that can be fixed or added to the alignment? Next look at conservation. Can you pull apart blocks that might be "overaligned" to discover motifs? Precomputed WAR alignments are attached below, or use your own.