function [names]=genefactor(A,s,Des,ngene,j)
%  using factor loading matrix A, factor j, correlation vector s, 
%  find the ngene genes most loaded on the factor, plot their 
%  correlations in s against factor loadings, and return their names
% 
  [a,i]=sort(abs(A(:,j))); i=flipud(i);  ind=i(1:ngene); 
  scatter(A(ind,j),s(ind)); 
  xlabel(['Loadings on factor',int2str(j)])
  ylabel('Correlation with status')
  names=Des(ind,:); 
%