Data Mining and Visualization of Mouse Genome Data:

Last edited by
Quratt ul ain Siddique

Summary:

This paper discusses  the data mining of  the genomics of  the  mouse  that  is  an  area  of  importance  because  of  its  relationship  to  understanding  of  basic  genetics  of  other mammals and in particular the human as well as livestock genetics and  its breeding.

The data mining  tools of multiplot, data partition,  clustering, self-organized maps  (SOM),  regression,  association,  and neural networks were all used in  this research The  paper  has  demonstrated  the  data  mining  and visualization  results  including  virtual  gene  map,  mouse genomic  features  on  chromosome,  clustering,  cluster proximity,  T-Scores  effect,  self-organizing  map,  and regression analysis. One of  the novelties of  this  research is that  the data mining  is performed at the genomic level of a mammal that is commonly used as prototype testings for humans.

The  data mining  performed  on  the mouse  genome  data indicated  a  linearity  of  regression  for  the  B16F0 Chromosone,  significant  reduction  in  the  average  error upon  using  neural  network  algorithms,  significant  effect in  the  visulization plots  upon using  self-organized maps (SOM),  and  a  nonlinear  relationship  of  the  cubicclustering criterion with  discontinities when  the  number of clusters reached 22 and 38.

The  results of data mining performed also  indicated  that it  was  useful  to  visualize  at  the  genomic  level  for  the mouse  data. The  analysis  shown  here  can  also  help researchers who are interested in genome data, and others to  visualize  the  use  of  data  mining  at  this  micro dimensional level.

Future  directions  of  the  research  are  to  continue  to perform  more  data  mining  of  the  mouse  genome  data. This  may  entail  using  other  data  mining  tools  and software.  Other  future  directions  are  to  perform  data mining  for  other  data  bases  such  as  for  other mammals that are of evolutionary  relationship  to humans, and also other genomic databases of differing dimensionalities  to contrast  the  findings  of  the  research  presented  in  this paper.