PureCN paper revision
2019-12-04
1 Setup
2 LUAD tumor-only mode
2.1 Load data
Results from 442 samples are plotted (out of 507 tumor-only processed samples)
script_dir = "/home/sehyun/Documents/github/PureCN_manuscript"
results_dir = file.path(script_dir, "luad/Results/purity_ploidy")
purecn_puri_ploi = readRDS(file.path(results_dir, "data/luad_ABS_w_tumor_only.rds"))
# 442 obs. which have matched normal samples
# paired = readRDS(file.path(results_dir, "data/luad_ABS_w_matching_normal.rds"))
# write.csv(paired$fullname, "~/Documents/github/PureCN_manuscript/luad/Results/duplication/luad_442.csv")
luad_442 = read.csv("~/Documents/github/PureCN_manuscript/luad/Results/duplication/luad_442.csv")[,2]
purecn_puri_ploi = purecn_puri_ploi[which(purecn_puri_ploi$fullname %in% luad_442),]2.2 Purity and Ploidy Comparisons
These are Figure 1C and 1D in the manuscript.
3 Purity outliers
Top 80% of purity-concordant samples are marked in red and the others are colored in cyan. If I calculate Pearson correlation only with the concordant samples, it give 0.98.
3.1 Ploidy of purity outliers
[Left] Ploidy of PureCN and ABSOLUTE results from 20% of purity-discordant
samples (the cyan dots from the above plot) are plotted. Pearson correltation
coefficient is 0.19.
[Right] Ploidy of PureCN and ABSOLUTE results from the purity-concordant
samples. Pearson correltation coefficient is 0.65.
–> So a wrong purity estimates lead to a wrong ploidy estimates?
4 Ploidy outliers
Top 80% of ploidy-concordant samples are marked in red and the others are colored in cyan. If I calculate Pearson correlation only with the concordant samples, it give 0.97.
4.1 Purity of ploidy outliers
[Left] Purity of PureCN and ABSOLUTE results from 20% of ploidy-discordant
samples (the cyan dots from the above plot) are plotted. Pearson correltation
coefficient is 0.69.
[Right] Purity of PureCN and ABSOLUTE results from the ploidy-concordant
samples. Pearson correltation coefficient is 0.87.
5 Extreme cases (2%)
5.1 Purity outliers
puriOutlierInd = which(df_puri_outlier$absdiff == "discordant")
df_puri_outlier[puriOutlierInd,]## SampleId Purity_ABS Ploidy_ABS fullname
## 30 TCGA-05-5425-01 0.71 1.08 TCGA-05-5425-01A-02D
## 130 TCGA-50-5930-01 0.95 2.08 TCGA-50-5930-01A-11D
## 143 TCGA-50-6591-01 0.90 4.67 TCGA-50-6591-01A-11D
## 150 TCGA-50-8457-01 0.60 2.03 TCGA-50-8457-01A-11D
## 177 TCGA-55-6984-01 0.81 4.06 TCGA-55-6984-01A-11D
## 328 TCGA-78-7153-01 0.57 4.00 TCGA-78-7153-01A-11D
## 381 TCGA-86-8671-01 1.00 2.00 TCGA-86-8671-01A-11D
## 424 TCGA-97-7546-01 0.67 1.85 TCGA-97-7546-01A-11D
## 454 TCGA-99-AA5R-01 0.96 2.00 TCGA-99-AA5R-01A-11D
## Purity_tumor_only Ploidy_tumor_only Flagged Comment capture_kit
## 30 0.30 2.827928 FALSE <NA> 931070
## 130 0.34 2.122316 FALSE <NA> 931070
## 143 0.43 2.677163 FALSE <NA> 931070
## 150 0.16 2.129037 TRUE LOW PURITY 931070
## 177 0.41 2.054388 TRUE POOR GOF (79.1%) 931070
## 328 0.24 2.356133 TRUE LOW PURITY 931070
## 381 0.17 1.944552 TRUE LOW PURITY 931070
## 424 0.24 2.143490 TRUE LOW PURITY 931070
## 454 0.34 2.003381 TRUE NON-ABERRANT 931070
## absdiff
## 30 discordant
## 130 discordant
## 143 discordant
## 150 discordant
## 177 discordant
## 328 discordant
## 381 discordant
## 424 discordant
## 454 discordant5.2 Ploidy outliers
ploiOutlierInd = which(df_ploi_outlier$absdiff == "discordant")
df_ploi_outlier[ploiOutlierInd,]## SampleId Purity_ABS Ploidy_ABS fullname
## 15 TCGA-05-4410-01 0.24 4.49 TCGA-05-4410-01A-21D
## 101 TCGA-49-4512-01 0.32 4.56 TCGA-49-4512-01A-21D
## 134 TCGA-50-5935-01 0.39 5.81 TCGA-50-5935-01A-11D
## 265 TCGA-64-1681-01 0.31 5.63 TCGA-64-1681-01A-11D
## 280 TCGA-67-6217-01 0.25 6.17 TCGA-67-6217-01A-11D
## 400 TCGA-91-7771-01 0.32 5.00 TCGA-91-7771-01A-11D
## 413 TCGA-95-7039-01 0.41 5.66 TCGA-95-7039-01A-11D
## 428 TCGA-97-7554-01 0.29 5.21 TCGA-97-7554-01A-11D
## 465 TCGA-L9-A743-01 0.09 6.48 TCGA-L9-A743-01A-43D
## Purity_tumor_only Ploidy_tumor_only Flagged Comment capture_kit
## 15 0.25 2.304502 TRUE EXCESSIVE LOH 931070
## 101 0.43 2.289707 FALSE <NA> 931070
## 134 0.52 3.171802 FALSE <NA> 931070
## 265 0.39 3.158382 FALSE <NA> 931070
## 280 0.33 2.026858 TRUE POOR GOF (79.4%) 931070
## 400 0.31 2.286902 TRUE POOR GOF (78.7%) 931070
## 413 0.43 3.024252 FALSE <NA> 931070
## 428 0.30 2.846310 FALSE <NA> 931070
## 465 0.16 1.750916 TRUE EXCESSIVE LOH 931070
## absdiff
## 15 discordant
## 101 discordant
## 134 discordant
## 265 discordant
## 280 discordant
## 400 discordant
## 413 discordant
## 428 discordant
## 465 discordant