Assignments for Today:

Read http://en.wikipedia.org/wiki/BLAST and http://en.wikipedia.org/wiki/FASTA
(read through, you don't need to recall the formulas, but you should understand the principle - )
Read through http://en.wikipedia.org/wiki/FASTA_format
http://en.wikipedia.org/wiki/Substitution_matrix and here (optional) for a curious twist
Read through the entry on the Bonferroni correction on wikipedia: http://en.wikipedia.org/wiki/Bonferroni_correction (a concise version is here, a discussion of fishing expeditions is in the introduction here. Especially this link is worth noting. Optional: Article in the New Yorker).

Assignments for Wednesday

Takehome exam #2 is due
Read http://en.wikipedia.org/wiki/Standard_score
Understand the difference between false positives and false negatives (see error types )

Discussion of Take Home Exam 1 (w anwers) (Excel file for extra question) -- Pan Genome and KS Plot slides here

DataBank Searches

Sequence and structure databanks can be divided into many different categories.
One of the most important is:

Supervised databanks with gatekeeper.

Examples:

Swissprot

Refseq (at NCBI)

Entries are checked for accuracy.
+ more reliable annotations
-- frequently out of date

Repositories without gatekeeper.

Examples:

GenBank

EMBL

TrEMBL

Everything is accepted.
+ everything is available
-- many duplicates
-- poor reliability of annotations

One problem in maintaining databanks (supervised and aunsupervised) is "owner ship" of sequences, which in many databanks prevents a continuous update of sequences. Even if errors are detected, they are not easily removed form the databank. E.g. ATP synthase operons in E.coli see http://mic.sgmjournals.org/cgi/content-nw/full/156/7/1909/F1

Even species names are often wrongly assigned (slides)

Types of Error in a Databank search

False positives: The number of false positives are estimated in the E-value. The P-value or significance value gives the probability that a positive identification is made in error (same as with drug tests).
Danger: avoid fishing expiditions. If you do 100 tests on random data, you expect one to be positive at the 1% significance level.

You could apply the Bonferroni correction:

The significance level for the individual test is calculated through dividing the overall desired significance level by the number of parallel tests. The hypothesis to be be rejected is that Not all of the individual tests are significantly different from chance. (all in the sense of "at least one"

False negatives: Homologous sequences in the databank that are not recognized as such. If there are only 12000 different protein families, an average a sequence should have (size of the databank)/12000 matches. In other words, the number of false negatives is probably very large.

Discussion: Decay of significance. Can this be corrected?

Addendum to virus and life discussion:

Cellular automata: A'life; John Conway's game of life. [rules: a cell survives if it has two or three living neighbors. A new cell is created on a "dead" square if it has exactly three living neighbors.] The game was popularized by Martin Gardner in Scientific American in 1970.

Examples:

http://www.pbs.org/opb/intimatestrangers/game/index.html
http://www.mindspring.com/~alanh/life/ (this lets you specify living cells)

More information on digital life is at Digital evolution homepage at MSU.
Karl Sims' virtual creatures are worth a look, movie here.
A similar approach to evolution in silico is here.

A'life not really alive

Go through coral of life ppt slides

What does the term phylogeny mean? Is it compatible with reticulation events?