| Wiki Markup |
|---|
{color:#ff0000}{_}Important update (January 20th, 2011): the data below have been corrected for the BCR batch which is not necessarily the processing batch. The dataset needs to be reanalyzed._{color}_ _
h4. Correlation between BCR batch and the processing batch for 27k arrays (January 20, 2012):
{csv}Batch on the download page,"# after ""HumanMethylation27k"" in the file name, Level 1",Batch as the sixth field in the patient barcode
Batch 1,2,0186
Batch 2,3,0199
Batch 3,4,0218
Batch 4,no data,
Batch 5,no data,
Batch 6,no data,
Batch 7,no data,
Batch 8,no data,
Batch 10,1,0392
Batch 16,5,0521
Batch 20,6,0595
Batch 26,7,0788
Batch 38,8,0915
Batch 62,9,1228
Batch 79,no data,
Batch 111,no data,
Batch 130,no data,
Batch 174,no data,{csv}
h5. Batch vs clinical traits
Number of clinical traits: 31
Number of batches based on DNA methylation data: 19
Relationship between batch and the center:
{code:collapse=true}> table(batchID,two)
two
batchID 02 06 08 12 14 15 16 19 26 27 28 32 41 74 76 81 87
0186 25 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0199 17 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0218 0 28 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0242 0 53 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0279 34 17 15 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0287 26 16 15 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0297 13 0 19 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0314 2 7 6 5 0 0 0 0 0 0 0 0 0 0 0 0 0
0337 0 10 0 16 0 1 0 0 0 0 0 0 0 0 0 0 0
0392 0 7 0 10 6 0 5 0 0 0 0 0 0 0 0 0 0
0521 0 4 0 12 12 3 7 6 3 0 0 0 0 0 0 0 0
0595 0 4 0 4 13 0 1 10 1 4 9 0 0 0 0 0 0
0788 5 12 0 0 3 0 0 0 0 13 10 4 0 0 0 0 0
0915 0 0 0 9 2 0 0 9 0 0 0 5 4 0 0 0 0
1228 0 1 0 0 2 0 1 3 1 0 0 15 4 0 0 0 0
1481 0 5 0 3 0 0 0 0 6 0 11 1 0 0 9 0 0
1697 0 13 0 0 1 1 0 10 1 0 3 2 1 0 5 1 1
1844 0 10 0 0 5 0 0 1 2 0 1 0 1 6 10 1 0
2004 0 7 0 0 3 0 0 4 0 0 0 0 0 0 0 0 0{code}
Relationship between batch and clinical variable, significant correlations (entire table can be found [here|^BatchClinicalInfoCorrelationsGBM.txt])
{csv}GBM_clinical,DataType,NumberOfNAs,Test,Pvalue
year_of_initial_pathologic_diagnosis,integer,35,Kruskal-Wallis rank sum test,1.34E-64
pretreatment_history,factor,35,Pearson's Chi-squared test,2.98E-29
histological_type,factor,35,Pearson's Chi-squared test,6.11E-20
initial_pathologic_diagnosis_method,factor,37,Pearson's Chi-squared test,5.24E-15
vital_status,factor,36,Pearson's Chi-squared test,8.36E-15
hormonal_therapy,factor,59,Pearson's Chi-squared test,6.50E-14
targeted_molecular_therapy,factor,65,Pearson's Chi-squared test,5.80E-10
additional_pharmaceutical_therapy,factor,72,Pearson's Chi-squared test,4.19E-05
additional_drug_therapy,factor,73,Pearson's Chi-squared test,5.08E-05
days_to_last_followup,integer,35,Kruskal-Wallis rank sum test,2.90E-04
person_neoplasm_cancer_status,factor,88,Pearson's Chi-squared test,4.81E-04
additional_chemo_therapy,factor,106,Pearson's Chi-squared test,5.18E-03
days_to_death,integer,169,Kruskal-Wallis rank sum test,6.30E-03
days_to_birth,integer,35,Kruskal-Wallis rank sum test,1.13E-02
age_at_initial_pathologic_diagnosis,integer,35,Kruskal-Wallis rank sum test,1.20E-02{csv}
h5. Survival vs batch
Code for automatic analysis of survival and correlation with clinical traits can be found [here|^SurvivalBasicAnalysis.R].
Kaplan Meier Curve and survival by batch:
!KaplanMeierCurveGBM.png|thumbnail! !SurvivalByBatchGBM.png|thumbnail!
Summary of the cox proportional hazards model can be found [here|^SurvivalBatchSummaryStatisticsGBM.txt], batch shows significant correlation with survival (Likelihood ratio test= 31 on 17 df, p=0.02; Wald test = 28.17 on 17 df, p=0.04297; Score (logrank) test = 29.48 on 17 df, p=0.03035)
I need to state here (and for all cancer types from TCGA that I have analyzed and will analyze) that the p-values for the association of the batches with the clinical traits correspond to ALL batches. However, actual DNA methylation (or other data) may not be available for all batches yet. For example, for GBM I have 9 batches for the downloaded 286 patients. I still see significant correlation between these batches and the clinical traits (p values might be bigger though. For example, correlation between histological type and batches is 3.9e-11).
h5. DNA methylation
Downloaded the data in the last week of December, 27k, Level1, 294 patients. Weird format for the files, methylated and unmethylated probe intensities are in the first and second columns, different from the format that was used for other datasets.
Compared list of DNA methylation patients with the technical info, tech info is available for only 286 patients. Stick with those for the analysis. Didn't split into the subsets of "green" and "red" probes.
!GBM_Mvalue_unnorm_distrib.png|thumbnail! !GBM_Mvalue_unnorm_relativeVariance.png|thumbnail! !GBM_Mvalue_noNorm_PC1outliers.png|thumbnail!
Technical variables:
{code:collapse=true}#Correlation of DNA methylation (real data) batches and the center:
> table(two,shortMeth$batchID)
two 0186 0199 0218 0392 0521 0595 0788 0915 1228
02 25 17 0 0 0 0 5 0 0
06 0 0 20 7 4 4 12 0 1
12 0 0 0 10 12 4 0 9 0
14 0 0 0 6 12 13 3 2 2
15 0 0 0 0 3 0 0 0 0
16 0 0 0 5 7 1 0 0 1
19 0 0 0 0 6 10 0 9 3
26 0 0 0 0 3 1 0 0 1
27 0 0 0 0 0 4 13 0 0
28 0 0 0 0 0 9 10 0 0
32 0 0 0 0 0 0 4 5 15
41 0 0 0 0 0 0 0 4 4
#Concentration can be probably treated as a continuous variable:
> shortMeth[1:10,3]
[1] 0.15 ug/uL 0.15 ug/uL 0.140 ug/uL 0.14 ug/uL 0.14 ug/uL 0.140 ug/uL
[7] 0.14 ug/uL 0.189 ug/uL 0.167 ug/uL 0.164 ug/uL
81 Levels: 0.071 ug/uL 0.12514558 ug/uL 0.134 ug/uL ... .19 ug/uL
> table(shortMeth$plate_column)
1 2 3 4 5 6
66 68 64 43 24 21
> table(shortMeth$plate_row)
A B C D E F G H
37 36 37 38 37 35 35 31
> table(shortMeth$shortDay)
13-5-2009 13-9-2010 14-4-2010 18-9-2008 20-1-2010 24-8-2009 29-6-2007 3-5-2007
47 27 29 28 47 46 20 17
4-4-2007
25
#Also, day of shipment is the same as the batch which I have seen before
> table(shortMeth$batchID,shortMeth$shortDay)
13-5-2009 13-9-2010 14-4-2010 18-9-2008 20-1-2010 24-8-2009 29-6-2007
0186 0 0 0 0 0 0 0
0199 0 0 0 0 0 0 0
0218 0 0 0 0 0 0 20
0392 0 0 0 28 0 0 0
0521 47 0 0 0 0 0 0
0595 0 0 0 0 0 46 0
0788 0 0 0 0 47 0 0
0915 0 0 29 0 0 0 0
1228 0 27 0 0 0 0 0
3-5-2007 4-4-2007
0186 0 25
0199 17 0
0218 0 0
0392 0 0
0521 0 0
0595 0 0
0788 0 0
0915 0 0
1228 0 0
{code}
Correlations with the first principal components:
{code:collapse=true}batchID amount concentration plate_column plate_row shortDay
V1 2.675388e-34 2.305474e-24 0.0000837995 0.002205121 0.33206728 2.675388e-34
V2 3.542743e-02 9.080686e-02 0.8088404440 0.160738153 0.28798646 3.542743e-02
V3 1.232019e-01 9.583923e-02 0.2891811969 0.711142850 0.04714336 1.232019e-01
V4 2.845743e-01 2.493076e-01 0.2479220305 0.353589457 0.45597681 2.845743e-01
V5 2.697899e-02 2.963406e-02 0.4862742532 0.956135668 0.51387714 2.697899e-02
V6 1.639557e-39 6.765313e-13 0.0005379532 0.397065857 0.73169940 1.639557e-39{code}
Converting concentration to a continuous variable and correlating it with the principal components showed that it is correlation with PC1 (p-value = 1.117e-06) but not with PC2 (p-value = 0.5189).
Begin by removing batch:
!GBM_Mvalue_batchRemoved_RelativeVariance.png|thumbnail!
Correlation with technical variables:
{code:collapse=true}batchID amount concentration plate_column plate_row shortDay
V1 0.9078327 0.8534161 0.4472358 0.35474247 0.01031398 0.9078327
V2 0.9999772 0.9989092 0.7858814 0.05639337 0.35637921 0.9999772
V3 0.9729759 0.6323455 0.2021242 0.66153708 0.05678137 0.9729759
V4 0.9999986 0.9803822 0.3390278 0.26653534 0.40375589 0.9999986
V5 0.9999998 0.9891846 0.7570124 0.86562477 0.46551409 0.9999998
V6 0.9999999 0.9830042 0.6796671 0.25051874 0.95548597 0.9999999{code}
In addition, removing batch also removed correlation with the concentration (p value was calculated by treating the concentration as a continuous variable). It seems that there is still some correlation with the plate row (PC1). Remove batch and the plate row:
!GBM_Mvalue_batchRowRemoved_RelativeVariance.png|thumbnail! !GBM_Mvalue_BatchRowRemoved_PC1outliers.png|thumbnail!
It looks like there is slight decrease in the relative variance after both variables. Also, I swear\! Unnormalized data looks better than normalized. Correlation with the tech variables:
{code:collapse=true}batchID amount concentration plate_column plate_row shortDay
V1 0.9783321 0.8868898 0.4505665 0.27982900 0.9995098 0.9783321
V2 0.9999991 0.9990549 0.7543368 0.04823674 0.9998402 0.9999991
V3 0.9930916 0.7217562 0.2847028 0.68888360 0.9906491 0.9930916
V4 0.9999936 0.9967183 0.3954235 0.24689131 0.9999548 0.9999936
V5 0.9999911 0.9685936 0.7412973 0.95834136 0.9999678 0.9999911
V6 0.9999999 0.9905920 0.6786770 0.22633527 1.0000000 0.9999999{code}
In addition, after removing the batch and the plate row I tested for correlation with the center: no correlation (p-value = 0.2305).
Accept this normalization. ExpressionSet is available. |
Content Comparison
General
Content
Integrations
App links