"NA" in putativeSNV.csv$LDrefine_trioLoci_score and LDrefine_merged_score

[Arsenalwins]

Thank you for the great pipeline! Following your intrinsic instructions, I have successfully run through the fantastic pipeline for Somatic SNV calling from scRNA-seq. However, I found that the cellScan step output is slightly different from your tutorial, and there are issues with the putativeSNV.csv

(base) [cpuddesperadio@sl-c0032 working]$ python ${path}/src/Monopogen.py somatic \
    -a ${path}/apps -r region.lst -t 1 \
    -i bm -l /lustre/home/cpuddesperadio/CSC/findcsc/check-by-0.5/sample54.csv -s cellScan \
    -g /lustre/home/cpuddesperadio/Monopogen/working/hg38.fa

[2024-10-19 09:33:49,064] INFO     Monopogen.py Collect single cell level information from sequencing data...
170157:108094:26219:26680:1119:66959
[2024-10-19 10:29:24,543] INFO     Monopogen.py Success! See instructions above.

When I finished the pipeline, I found that putativeSNV.csv had LDrefine_trioLoci_score and LDrefine_merged_score
all as NA, and there are also some NAs in LDrefine_twoLoci_score. None of the SNVs pass the criteria for visualization. I wonder what's happened and what should I do? Thank you again for the great tool!谢谢！
chr2.putativeSNVs.csv
svm_feature.chr2.pdf
LDrefinement_germline.chr2.pdf

[jinzhuangdou]

Based on the output, it appears that the sequencing depth in your data is quite low, resulting in insufficient germline SNVs to serve as anchors for identifying somatic SNVs. You might consider trying our new update, which still provides the cell x SNV matrix even if the LD refinement step fails. You can still perform some cell type level filtering based on such output.