Name (abbrev)

Name (full)

Category

Last update

QSARs

Learning Qualitative Structure Activity Relationships (QSARs) for Pyrimidine and Triazine Compounds

Molecular Biology

b D, Y

Application domain

Further specifications

Learning Qualitative Structure Activity Relationships (QSARs) for Inhibition of E. Coli Dihydrofolate Reductase

two datasets (pyrimidines and triazines)

Type

Format

Complexity

ILP

Golem

3.5 MB

WWW / FTP

ftp://ftp.mlnet.org/ml-archive/ILP/public/data/drug

Contact person(s)

Related group(s)

Optional contact address

2. University of York, Dept of CS

References

King R.D., Muggleton S., Sternberg M.J.E. (1992)
Drug design by machine learning: The use of ILP to model
the structure-activity relationships of trimethoprim analogues
binding the dihydrofolate reductase, Proc. of Nat. Academy of
Sciences, 89 (23): 11322-11326
King R.D., Muggleton S., Sternberg M.J.E. (1995) Relating chemical activity to structure: an examination of ILP successes, New Gen. Computing (to appear)

Annotations

Structure activity relationships (SAR) describe empirically derived relationships between the chemical structure and the activity of considered drugs. In a typical SAR problem a set of chemicals of known stucture and activity are given, and the goal is to construct a predictive theory relating the structure of a component to its activity. Both available datasets described in [King 92, King 95] concern the classic drug design problem of inhibition of E. Coli Dihydrofolate Reductase by pyrimidines and triazines.

Pyrimidine compounds are antibiotics based on a common template to which chemical groups can be added at 3 possible substitution positions. A chemical group is an atom or a set of structurally connected atoms (that can be substituted together as a unit) characterized by well defined chemical properties. Some of these properties are encoded within the background knowledge as facts, e.g. polar(br,polar3) states that a bromin atom has a polarity 3. Different considered pyrimidine compounds are identified by the content of the considered 3 positions. Positive examples represent pairs of drugs, the activity of one being known to be higher then that of the other one.

Triazines act as anti-cancer agents by preferentially inhibiting reproducing cells. Like pyrimidines, they have a common template but this one is much more complicated with 5 possible substitution positions. The research goal is to identify the rules for predicting activity of different derived compounds. Formally, this task and its structure is closely related to the pyrimidine problem.

Comments

Two datasets are given: pyramidines and triazines. Both of them can be freely distributed as long as appropriate references are given to the papers mentioned.