Support Vector Machine GPCR Subfamily Classification Results for NP_000666

>NP_000666 MPIMGSSVYITVELAIAVLAILGNVLVCWAVWLNSNLQNVTNYFVVSLAAADIAVGVLAIPFAITISTGF CAACHGCLFIACFVLVLTQSSIFSLLAIAIDRYIAIRIPLRYNGLVTGTRAKGIIAICWVLSFAIGLTPM LGWNNCGQPKEGKNHSQGCGEGQVACLFEDVVPMNYMVYFNFFACVLVPLLLMLGVYLRIFLAARRQLKQ MESQPLPGERARSTLQKEVHAAKSLAIIVGLFALCWLPLHIINCFTFFCPDCSHAPLWLMYLAIVLSHTN SVVNPFIYAYRIREFRQTFRKIIRSHVLRQQEPFKAAGTSARVLAAHGSDGEQVSLRLNGHPPGVWANGS APHPERRPNGYALGLVSGGSAQESQGNTGLPDVELLSHELKGVCPEPPGLDDPLAQDGAGVS



Here is a list of Class A Rhodopsin like subfamilies with support vector machine models and the score this sequence received with respect to each model.

1.2525487 Nucleotide-like GPCRDB Nucleotide-like
  1.0815737 Adenosine GPCRDB Adenosine
   1.348909 Adenosine type 2 GPCRDB Adenosine type 2
  -0.9089301 Adenosine type 3
  -0.9974859 Adenosine type 1
 -1.3468437Purinoceptors
-0.8123337Amine
-0.8977933Olfactory
-0.98343337(Rhod)opsin
-0.99418974Prostanoid
-1.0016372Gonadotropin-releasing hormone
-1.002151Cannabis
-1.0075986Peptide
-1.0343342Thyrotropin-releasing hormone & Secretagogue
-1.0572089Class A Orphan/other
-1.086083Melatonin
-1.1239582Lysosphingolipid & LPA (EDG)
-1.128787Hormone protein
-1.1500611Viral
-1.1983904Platelet activating factor




HOW TO INTERPRET THE SCORES:

Subfamily Scoring:

The support vector machine for each subfamily is trained to score positive examples (members of the subfamily) as 1.0 and negative examples (non-members) as -1.0

If this sequence receives a negative score with respect to a subfamily, it is probably not a member of the subfamily.

If this sequence receives a positive score with respect to a subfamily, it is probably in the subfamily.

In general, the score's distance from zero gives you the confidence level of the prediction. Positive scores greater than +1.0 mean a classifier has strongly accepted your sequence. Negative scores less than -1.0 mean a classifier has strongly rejected your sequence. For a more detailed evaluation of scores and confidence levels, take a look at Classification Statistics.
SOME ADDITIONAL INFORMATION:

There were 19 Class B Secretin like and 11 Class C Metabotropic glutamate / pheromone and 2 Class D Fungal pheromone and 5 Frizzled/Smoothened family and 4 Nematode chemoreceptors and 4 Vomeronasal receptors (V1R & V3R) and 73 Class A Rhodopsin like subfamilies whose classifiers were not run, because the families or subfamilies that contain them were rejected by classifiers higher in the hierarchy. For example, if the Amine receptor classifier scores this sequence negatively, it will not be checked with respect to the subfamilies of Amine receptors (Histamine, Serotonin, Dopamine, Octopamine and Acetylcholine (muscarinic) receptors).

Not all subfamilies have models yet. This sequence may be in a subfamily that has not yet been modeled.

Please cite: R. Karchin, K. Karplus and D. Haussler "Classifying G-Protein Coupled Receptors with Support Vector Machines" Bioinformatics 2002 in press [postscript, pdf]