Reconstructing Putative "Full Sequence" From Incomplete Contigs Using Available Complete Sequences As Target

Question

[Raghul]

Hi to all,

I have many contigs of rRNA which are small & incomplete. I have nearby relatives with complete sequences. I want to align all my sequences with this complete sequence as a target & reconstruct the possible full length rRNA with gaps. Multiple sequence alignment programs always tend to cluster & give poor alignments. So anybody with similar experience can give me ideas, so as I can do this better!

thank u raghul

Answer 1

[Hamish]

As you have noticed multiple sequence alignment tools do not work for this case, since your sequences do not contain a mutual overlap (one of the core assumptions for global multiple sequence alignment).

So you will need to look at alternative alignment methods, for this these would breakdown into three rough classes:

1. Guided sequence assembly. Using a reference sequence to assemble many smaller sequence reads.
2. Sequence mapping programs. These align a set short sequences to a longer sequence.
3. Sequence clustering programs. Create longer reads or pseudo-transcripts by clustering the sequences for a source.

Typical programs that come to mind are:

• Sequence assembly:
  • Staden
• Mapping tools:
  • EMBOSS est2genome part of the EMBOSS package.
  • GLSEARCH, a global in the query and local in the subject pairwise alignment tool, part of the FASTA suite
  • SSAHA tools, see http://www.sanger.ac.uk/resources/software/#t_assembly
  • sim4
• Sequence clustering:
  • CD-HIT

Many more options can be found on these Wikipedia pages:

• "Sequence assembly"
• "List of sequence alignment software"
• "Sequence clustering"

I suggest trying various programs with your set of sequences to see which work best for your case.