Entropy Analysis of QC Samples 7-20-12
2015-01-13azim58 - Entropy Analysis of QC Samples 7-20-12
Josh and I were discussing how necessary it is to normalize sample data
before calculating the entropy. I argued that the data does not need to
be normalized since any shifts in intensity will not affect the spread of
the distribution which is what effects the entropy value. Josh said we
could test how consistent the entropy values are on the QC samples which
should essentially be the same even though there may be some batch
differences.
Location of QC Data:
S:\Research\labdata\jaricher\QC Data\sera
and
S:\Research\labdata\jaricher\QC Data\biotin
I'll calculate the entropy of these samples
Results Placed here:
L:\storage\CIM Research Folder\DR\2012\7-23-12\qc entropy analysis
===========================================================================
Josh also mentioned a possible method of preparing data for normalization
using the concept of Bayesian Information Criteria (BIC). From what I
remember him saying, this method splits the distribution into groups so
in some ways it is kind of like a K means clustering theory.
===========================================================================
The entropy of the QC samples can vary drastically between batches over
time. Therefore, I may have been incorrect when I previously argued that
it is not important to calculate the entropy with normalized data since
entropy deals with the spread of intensities rather than degree of the
intensities. Calculating the entropy with normalized data may be
important to do.
Analysis and graphs can be found here
"L:\storage\CIM Research Folder\DR\2012\7-23-12\qc entropy
analysis\Entropy of QC Samples.pptx"
Original excel sheets here
"L:\storage\CIM Research Folder\DR\2012\7-23-12\qc entropy
analysis\biotion\Entropy values of biotin samples.xlsx"
and here
"L:\storage\CIM Research Folder\DR\2012\7-23-12\qc entropy
analysis\sera\Entropy values of Sera Samples.xlsx"