Section 9-4: Comparison Programs

[ Previous chapter ][ This chapter ][ Next chapter ] Programs may distinguish between the

Global Alignment: end-to-end alignment

Best LOCAL Homology: finding the best fragment in two sequences

Subsection 9.4.1

Two Sequences of Similar Length

Best suited for comparing homologous sequences from different species, or similar sequences with approximately the same length:

Subsection 9.4.2

Two Sequences of Different Length

Best suited for comparing sequences discovered in searches, or sequences with site homology rather than integral similarity.

Subsection 9.4.3

DNA and Protein Sequences

Best suited for comparing sequences discovered in searches, or sequences with site homology and a suspicious reading frame shift.

% framealign

Subsection 9.4.4

Programs to Display Two Aligned Sequences, Text

publish can display alignments (DNA or protein) in formatted fashion.

Subsection 9.4.5

Programs to Display Two Aligned Sequences, Graphics

Best suited for visualising overlaps or regions of homology. Needs Graphics - remember to have set the graphics environment with setplot correctly if you work with GCG locally. X-Windows setups have to set the DISPLAY environment correctly.

% bestfit -out

Next, display the graphics using the ".out" files generated by the commands given above.

Subsection 9.4.6

Significance Evaluation

Preparation of Data

comptable creates a comparison table if required. To create an own table, use a simplification file as described below . Comparison tables are PAM-250 (evolutionary matrix) by default, an other option is the alignment-generated matrix called genmoredata:blosum62.cmp.

shuffle randomises a sequence to allow significance estimates: Shuffle the sequences and retry the alignment. If the same or similar result scores as the previous are observed, you will get a malicious alignment biased by composition which cannot be aligned using the described methods.

Randomisation during Alignment

Best suited for estimating whether the alignment produced is (statistically) significant. Should be used with significantly more than the default 10 randomisation's (try at least 50).

% gap -ran=50

% bestfit -ran=50

================================= Begin Exercise 10

Pairwise sequence analysis: Understand the use of comparison matrices in the alignment procedure of protein sequences. Apply different algorithms to the sequences obtained from DNA after translation, and evaluate significance of the result on both DNA and Protein level.

Using previous exercise results, you should have two DNA sequences by now: my1.seq as the typed-in sequence, and my2.seq as the reading-frame extracted DNA sequence from the seqed exercise. You shall compare these two sequences now on DNA and protein level. If you haven't translated the sequences to protein level already, you should do this now.

To solve this problem, follow this schedule:

Set up the gcg environment and plotting environment if required (non-WPI users only).

Comparing the methods: Compare the typed-in sequence and the sequence retrieved from the database as obtained in the previous exercises on DNA level with gap and bestfit . Does strand orientation matter?

Compare the resulting protein sequences with gap and bestfit . What would be the result of a reading frame shift?

Creation of own tables: Generate an own comparison table ( comptable ) for protein alignments based on a simplification table which you can obtain with the command fetch simplify.txt .

Using own tables: Align the two protein sequences using gap with your own and the standard table. What is the difference in the output?

Result presentation: Correlate the result of the protein alignment with the DNA alignment, and display the DNA alignment with one peptide sequence using the 'publish' program.

================================= End Exercise 10

[next page] , or [overview] , or [table of contents]