Finding peptides that increased in SMC1fs relative to naive 5-5-12

2015-01-13

azim58 - Finding peptides that increased in SMC1fs relative to naive
5-5-12

some files used in analysis
L:\storage\CIM Research Folder\DR\2012\5-5-12\smc1fs anal
329 peptide list
"L:\storage\CIM Research Folder\DR\2012\5-5-12\smc1fs anal\329 peptides
in blocking experiment.xlsx"

The 329 peptides were selected by choosing the peptides that increased
2.5 fold from naive to smc1fs. This was done as follows as follows:
Genes selected from condition Sample Name 29P05681_RabbitSMC1.gpr that
have Normalized Data values that are greater than those in condition(s)
Sample Name 29P05687_IgG_NaiveRabbit.gpr by a factor of 2.5 fold.
Starting gene list: all genes. Naive is from 3-11-10 experiment and 1:500
SMC1fs is from 3-18-10 experiment
Note: This list can be found in Genespring at the following location:
Gene Lists->Kurt->Initial (Rabbit and some Mouse) SMC1
Immunosignatures->050412 SMC1fs Increasing
Analysis->SMC1fs_Increasing_2.5fold_050412

===========================================================================
Some heatmap info 5-7-12

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

After reading through a draft of my paper, Kathy asked why I was using
fold change and not using p values when I discussed this 329 peptide
list. I will see if I can find a similar list using p values.
One problem is that the 329 peptide list was constructed from two single
samples from two different experiments, and therefore no p value can be
listed.

I looked at the SMC1fs vs naive of the 122109 experiment. The TAFY,
TISKY, RVGEM, and AVSHQ peptides have fairly good p values and lower than
20% false discovery rates in this experiment, but they are not found in
the top 50 peptides. . .
see
L:\storage\CIM Research Folder\DR\2012\5-29-12\122109 exp anal
Those results were with normalized data. Maybe raw data will yield better
results?
The raw data did yield better results, but if I look at the 4 peptides
(TAFY. . .) then I see that I would have to include 400 peptides.

Maybe I can compare the naives from the 122109 experiment with the smc1fs
from the 3-18-10 experiment
I only have duplicates for the 1 to 100 dilution. comparison of the
naives to the 1 to 100 dilution on 122109 yields ideal results
When I combine the 1 to 500 dilution from the 3-18-10 experiment with the
1 to 500 dilutio from the 3-11-10 experiment, the results are not
terrible, but not ideal. Normalizing makes the data even worse.

I compared with the 041510 experiment, but the TISKY related peptides did
not rank very well (the TISKY peptide itself actually ranked #1 though)
L:\storage\CIM Research Folder\DR\2012\5-29-12\122109 exp anal\comparison
with 041510 exp

I compared with the 042210 experiment, and all of the TISKY peptides
ranked very well. It looks like this is the peptide list I'll use. There
are 108 peptides with a p value<0.0001.
C:\kurt\storage\CIM Research Folder\DR\2012\5-29-12\122109 exp
anal\comparison with 042210 exp
Now I need to get the intensities of these 108 peptides in the blocking
experiment and see how the peptides rank when seeing how they block with
the real antigen and not the negative control.
blocking experiment intensities here (all 100000 peptides
"C:\kurt\storage\CIM Research
Folder\DR\2012\6-5-12\intensities_in_blocking_experiment.txt"
raw intensities here
C:\kurt\storage\CIM Research Folder\DR\2012\6-5-12\raw intensities

108 SMC1fs peptides

Now I can make a new heatmap with these 108 peptides. This can be found
here:
L:\storage\CIM Research Folder\DR\2012\6-5-12\raw intensities\raw
intensity heatmap

azim58wiki: