Estimate of tumor proportion returns NA for most of the spots

[ElyasMo]

Thank you for the useful package.
I am trying to apply your method to my pediatric brain tumor data.
I have applied it to all 14 tissue sections at the same time by putting them in a tsv file as the tutorial for prostate cancer has suggested. These are not consecutive sections.

The results (loh_estimator_tumor_prop.tsv) show that most of spots received NA for the tumor proportion using the default parameters (almost 50% NA). I have tested different possibilities and it seems the parameters in the purity config file are the reason (if I try more relaxed parameters, e.g.,
secondary_min_umi : 200
tumorprop_threshold : 0.2
np_threshold : 0.7
np_eventminlen : 7
I get even more NA, close to 90% NA). Which parameters do you suggest to change to get more accurate results in this case.

Best,
Elyas

[Congm12]

Thank you for trying the tumor purity estimation module! The tumor purity estimation module heavily relies on B allele frequency at loss of heterozygosity regions, and if the regions are relatively small, it will be challenging to accurately estimate tumor purity in this way as opposed to using gene expression deconvolution methods. At this point, I suggest skipping the tumor purity estimation and directly running CalicoST assuming all spots are pure (by setting tumorprop_file to None). This run sometimes recognizes tumor-normal boundary region as a separate clone, but it should be possible to recognize this mixture by inspecting the spatial clone organization and the BAF-RDR plot (plots/rdr_baf_defaultcolor.pdf). And if it's okay to share the plot, I'd be happy to help with checking potential tumor-normal mixture clone, and/or check whether tumor purity estimation module can be further improved.

Best,
Cong