AS a continuing follow up the previous posts related to tuberculosis I thought you might find the following interesting: People have been working on a solution for some time...
JOURNAL OF COMPUTATIONAL BIOLOGY
Volume 13, Number 1, 2006
© Mary Ann Liebert, Inc.
Prediction of Membrane Proteins in
Mycobacterium tuberculosis Using a
Support Vector Machine Algorithm
"Mycobacterium tuberculosis (M.tb) infects one third of the world’s population and is the most
prevalent infectious disease, representing more than a quarter of the world’s preventable deaths.
Although drugs such as isoniazid and rifampin are used for treatment of tuberculosis, these drugs are
ineffective against multidrug-resistant TB (MDR-TB). It is thought that resistance to common antibiotics
and chemotherapeutic agents are conferred in MDR-TB by the unusual constitution of TB cell membrane,
which limits permeability. The inner layer of the outer membrane of M.tb contains large amounts of mycolic
acids, which are covalently linked to the peptidoglycan cell wall through arabinogalactan. This unique
cell envelope structure creates a barrier that limits permeation through the outer membrane by diffusive
mechanisms. Translocation across membranes in M.tb is reliant on membrane proteins as transporters,
porins, and channels. Consequently, to circumvent MDR-TB and generate new drugs capable of combating
MDR-TB infections, further insight into the unique constitution of M.tb membrane would be useful.
Additionally, information about membrane proteins that are essential to cellular function and permeability
in M.tb (e.g., channels, pumps, receptors) would be immensely helpful. This communication reports the
application of a machine-learning algorithm, Support Vector Machine (SVM; Joachims, 1999), to identify
M.tb membrane proteins. "
more info at the link if you want to read more...