Transcript DannyBLM2

Biological Language Modeling
Toolkit
“Graphing Utilities”
by: Danny Lam
Overview
• BLMT
Ex: Computes association measures in protein sequences
• Graphing Utilities
– Display how well the association measures or other
data (known or surmised) feature boundaries
• Step 1: Automatic extraction of feature boundaries from given
source files
• Step 2: Plot data along with feature positions along a sequence
BLMT : Mutual Information
• Mutual Information -> Computes
"mutual information”, which is a measure of
association between adjacent amino acids.
– Input: amino acid sequence file(s)
• (ex) Swiss prot SW datasets
– Output: file.mi.out.av ->
• first column is position in sequence
• second column is mutual information value
associated with that position
Feature Positions
• Extract feature position information (via
Swiss-prot)
– Extracellular (EC),
– Cytoplasmic (CP),
– Helices (H)
--> label where the EC, CP, and H regions are
in the sequence.
DR PROSITE; PS00238; OPSIN; 1.
KW Photoreceptor; Retinal protein; Transmembrane; Glycoprotein; Vision;
KW Phosphorylation; Lipoprotein; Palmitate; G-protein coupled receptor;
KW Acetylation; Retinitis pigmentosa; Disease mutation.
FT DOMAIN
1 36
EXTRACELLULAR.
FT TRANSMEM 37 61
1 (POTENTIAL).
FT DOMAIN
62 73
CYTOPLASMIC.
FT TRANSMEM 74 98
2 (POTENTIAL).
FT DOMAIN
99 113
EXTRACELLULAR.
FT TRANSMEM 114 133
3 (POTENTIAL).
FT DOMAIN
134 152
CYTOPLASMIC.
FT TRANSMEM 153 176
4 (POTENTIAL).
FT DOMAIN
177 202
EXTRACELLULAR.
FT TRANSMEM 203 230
5 (POTENTIAL).
FT DOMAIN
231 252
CYTOPLASMIC.
FT TRANSMEM 253 276
6 (POTENTIAL).
FT DOMAIN
277 284
EXTRACELLULAR.
FT TRANSMEM 285 309
7 (POTENTIAL).
FT DOMAIN
310 348
CYTOPLASMIC.
FT MOD_RES
1 1
ACETYLATION (BY SIMILARITY).
FT CARBOHYD 2 2
N-LINKED (GLCNAC...) (BY SIMILARITY).
FT CARBOHYD 15 15
N-LINKED (GLCNAC...) (BY SIMILARITY).
FT DISULFID 110 187
BY SIMILARITY.
FT BINDING 296 296
RETINAL CHROMOPHORE.
Problems/Solution
• Problems:
-Making one subplot graph (MATLAB) requires program
customization
- Generation of multiple subplots together requires more tedious
work. Waste of time and effort.
• Solution:
-Need clear interface to generate subplot graphs for you w/o
writing tedious matlab code.
[a1,b1]=textread(’test.out', '%d %f');
hold on
subplot(1,1,1);
hold on
hh1 = plot(a1, b1, 'linewidth',2.5);
hold on
ylabel('yaxis','fontsize',16, 'Color','k','fontweight','bold');
set(hh1, 'MarkerSize',5);
set(gca, 'YLim',[-1, 3]);
%set(gca,'ytick',[-.6,-.2,.2]
xdash = [NaN,62,73,NaN,134,152,NaN,231,252,NaN,310,348]; %cp
ydash = (-.2)*(ones(size(xdash)));
line(xdash,ydash,'color','y','linewidth',3);
xdash = [1,36,NaN,99,113,NaN,177,202,NaN,277,284,NaN]; %ec
ydash = (-.2)*(ones(size(xdash)));
line(xdash,ydash,'color','r','linewidth',3);
hold on
xlabel('x_axis','fontsize',16, 'Color','k');
print -dpsc -r0 sample;
Design Capabilities
• Access multiple mutual information output datasets
• Display combination of EC/CP/H position information on
MI datasets (color coded)
• Specify range (Y limits) and naming conventions (X
axis)
• Output into convenient picture files (ex: .tiff file).
Subplotter
• Version 1: (In house use only)
-Initially the program takes as input:
--> .SW file: (EC/CP/H)
--> .m file: (MATLAB file that code will
be generated in)
Subplotter ( Version 1)
***********************************
How many output files to textread: 1
What is the file to be textread into matlab program [output
file 1]: opsdh_1gpcr.out
How many TOTAL subplots do you request?: 1
************************************
Subplotter ( Version 1)
*********************
Subplot(1,1,1)
*********************
Which file do you want results to be graphed on this subplot?:
0: opsdh_1gpcr.out
Make selection (0): 0
++++++++++++++++++++++++++++++++++++++++++++++
How many items (EC,CP,H) do you want plotted
(1,2, 3: GPCR, 4: Loops)?:
+++++++++++++++++++++++++++++++++++++++++++++++
--> 3
Subplotter ( Version 1)
Specify Y-Axis Label? (y/n): n
Y-Axis Label: GPCR
Specify YLim? (y,n): n
Give name to X-Axis: sample
Give name to .tiff file for output (no extension!): sample
Matlab Program completed! wait ...
Subplotter (Version 1)
Current/Future Work
• Generate graphing utility for every tool on
the BLMT website.
Questions?