Work 071412

2015-01-13

azim58 - Work 071412

I had started making the findEverythingSimilarToI function. However, now
I think I want to remake the function using a slightly different
strategy. The first strategy involved looping through the similarity
scores to i, but then when a similar score was found the method would
branch off and compare everything to that method. Now I would rather
compare everything to i first, keep track of which things are similar,
and then branch off to those later.

initial code for first attempt of findEverythingSimilarToI function

===========================================================================
public ArrayList findEverythingSimilarToI(int i)
{
ArrayList things_similar_to_I = new ArrayList();
things_similar_to_I.add(i);

for(int j=0; j<((ArrayList)working_list.get(i)).size(); j++)
{
double current_value =
Double.valueOf(((ArrayList)working_list.get(i)).get(j).toString()).doubleVa
lue();
if(current_value!=-1)
{
if(current_value>threshold)
{
things_similar_to_I.add(i+j+1);
((ArrayList)working_list.get(i)).set(j, -1);
MotifGrouper2 mg = new Motif
}
}
}
}

===========================================================================

Made ItemData class.

1430 Finished making MotifGrouper2 class. Everything seems to be working.
In order to test it, I want to make one more fake score file, and see if
everything is grouped correctly.
My fake data scores could be represented in a matrix like this.
[-1, 10, 1, 1, 1]
[-, -1, 1, 50, 10]
[-, -, -1, 1,1]
[-, -, -, -1, 20]
[-, -, -, -, -1]

As an arraylist of arraylists this matrix would be represented like this:
[10, 1, 1, 1],[1, 50, 10],[1,1],[20],[]

My program should group items 0,1,3, and 4 with group 2 in it's own
group. The first group would be created all from the first iteration when
all similar items with 0 are found. The representative from the first
group should be item 3 since it has the highest score with everything.
The average score for 4 should be (1+50+20)/3.0 = 23.67.

I'm going to make quite a few changes to my
public ItemData findEverythingSimilarToItemI(int i)
method. Here's what it was before the changes
some code as of 071412_1500

1535 Everything in my program works as I expect it to now which means
that I can now group the motifs as well as choose the most representative
motif in the motif group.
I don't think I need the MotifGrouper class any longer
MotifGrouper Class as of 071412
Here is my MotifGrouper2 class which I plan to rename to MotifGrouper
L:\storage\CIM Research Folder\DR\2012\7-14-12
I also put the current versions of the MotifPeptideAnalysis and the
ItemData class there.

I think the next time I work on this I will want to use glam2scan to
assign the peptides to the different motif groups.

azim58wiki: