cluster.fragments

Hi, everyone.

I wonder whether the “cluster.fragments” commnad can replace the “cluster” command.

One trial using a small set of sequences and the “cluster.fragments” command with the option “percent=3” gave a better result than does the combination of “pairwise.seqs” and “cluster” or the combination of “dist.seqs” and “cluster”.

How do you think?

Also, I wonder what the alignment method is when the “cluster.fragments” command calculates the percent or difference(blastn, gotoh, or needleman?).

Thank you.

Hyungz.

The alignment method is needleman with gapOpen=-1, match=1, and mismatch=-1.

Thank you, westcott.

Hello,
I wonder what the result of cluster.fragments would be in situation when you have 1 short sequence that is identical to fragments of two longer sequences, however these too long sequences arent identical to each other?

Thanks

One trial using a small set of sequences and the “cluster.fragments” command with the option “percent=3” gave a better result than does the combination of “pairwise.seqs” and “cluster” or the combination of “dist.seqs” and “cluster”.

And what was your objective metric for making this assertion that cluster.fragment is better?