|
Sequence Homolog Searching
- Applications:
- Search your new sequence against a database.
- Search a known sequence against a new database or genome
to look for new relatives.
- Suggested Methods:
- BLAST Variants on BlastMachine
- BLASTN - fast and often
sufficient.
- TBLASTX - More tolerant of sequence variation and frame-shifts,
but slower
Remote Homolog Searching:
- Applications:
- Find distant relatives of a gene or protein.
- Suggested Methods:
- Smith-Waterman (GeneMatcher2) - More sensitive than
BLAST when comparing highly diverged sequences, but much slower.
- Double Frame Smith-Waterman (GeneMatcher2)- like
TBLASTX above, it is more tolerant of sequence variation and
frame-shifts. Very computationally intensive.
-
PSI-BLAST (BlastMachine) - Iteratively explore for
distant members of a protein family.
-
Fugu genome (avail. on BlastMachine
and GeneMatcher2) - The Fugu genome is very compact,
and is the most evolutionarily distant vertebrate from humans
that has been almost fully sequenced. It has been successfully
used as a bridge to find distant human family members of
a newly discovered human gene (internal Genome Center project).
- Because Profile HMMs are a consensus-type model, they may
be less useful for detecting very remote homologs than some
other methods.
EST Annotation
- Applications:
- Suggested Methods:
- BLASTN (BlastMachine).
- TBLASTX (BlastMachine) - for cross-species EST identification,
e.g. search non-human ESTs against the human or other gene
databases.
- Double-Affine Smith-Waterman - compare ESTs to genomic DNA.
Separate penalties for large insertions make it intron tolerant.
Protein Domain Detection
- Applications:
- Identify protein domains in a new sequence.
- Search for new members of a particular domain.
- Protein domain can imply function.
- Suggested Methods:
- Hidden Markov Models such as the pfam library (GeneMatcher2).
- Profile (PROSITE?)
Gene Finding
- Applications
- Search new or unfinished sequences for new genes.
- Suggested Methods
- Genewise (GeneMatcher2) - Genewise compares genomic
sequences to protein reference sequences or HMM models at
the protein level, allowing for introns (splicing) and frameshifts.
Optimal Alignment of Sequences
- Suggested Methods
- Smith-Waterman (GeneMatcher2).
Alignment of Short Sequences
- Suggested Methods
- Fasta - fasta lies between BLAST and Smith-Waterman in accuracy
and speed (on a given processor). It can be used to find
shorter matches than BLAST.
|