Sciclone clonevol integration

Question

[sm.hashemin]

Dear Chris, I tried using the integration of clonevol and sciclone (RUN) script. After infer.clonal.models I get

Sample 1: Tumor <-- Tumor
Sample 2: Relapse <-- Relapse
Using monoclonal model
Note: all VAFs were divided by 100 to convert from percentage to proportion.
Generating non-parametric boostrap samples...
Tumor : Enumerating clonal architectures...
Determining if cluster VAF is significantly positive...
Exluding clusters whose VAF < min.cluster.vaf=0.01
Non-positive VAF clusters:  
Tumor : 76 clonal architecture model(s) found

Relapse : Enumerating clonal architectures...
Determining if cluster VAF is significantly positive...
Exluding clusters whose VAF < min.cluster.vaf=0.01
Non-positive VAF clusters: 4,6 
Relapse : 2 clonal architecture model(s) found

Finding consensus models across samples...
Found  0 consensus model(s)
Found 0 consensus model(s)
Scoring models...
Pruning consensus clonal evolution trees....
Seeding aware pruning is:  off
Number of unique pruned consensus trees: 0

and then I get

f = generateFishplotInputs(results=res)
Error in 1:nrow(results$matched$index) : argument of length 0

Is it because there is no consensus models found between the two samples (tumor and relapse?

Best Mo

[GenoMax]

Tagging: Chris Miller

Answer 1

[Chris Miller]

Yes, if clonevol doesn't come up with a valid consensus model, then it's impossible to plot that model. You should examine your results carefully to identify 1) artifactual calls that may result in clusters that are wrong (one class of reference-problem-induced FP tends to be at about 15% in every sample). 2) look for outlier points that should be excluded because of CN or LOH (if you have points above 70%, something is wrong). 3) consider requiring higher depth for points to tighten up your clusters (depth of sequencing allowing!)