DarioS
commented
6 years ago I found exactly the same problem causes my analysis to fail. I suggest just to remove that gene from the IPS_genes.txt file and proceed. All of the other genes in the Treg category are present in GENCODE Genes, so I don't think it makes a big difference to the final score. Anyway, the software should be of better quality so it doesn't fail to produce any results if one gene symbol is missing. It's also quite messy and contains a lot of hard-coded values and hasn't ever been updated, That's a problem because it uses deprecated function that could soon be removed.
> head(warnings(), 3)
Warning messages:
1: `panel.margin` is deprecated. Please use `panel.spacing` property instead
2: `panel.margin` is deprecated. Please use `panel.spacing` property instead
3: `panel.margin` is deprecated. Please use `panel.spacing` property instead
dtiezzi
I'm working with TCGA RNAseq data (aligned to hg38) and realized there is no annotation for CCL3L1 gene in raw counts data from GDC repository. The corresponding ensembl id for CCL3L1 is ENSG00000277796, which is not available in TCGA data. However, I can find the ENSG00000276085 gene, which correspond to CCL3L3 gene. Looking at Ensemble website I realized the CCL3L1 and CCL3L3 correspond to similar location at the genome (http://www.ensembl.org/Homo_sapiens/Gene/Summary?g=ENSG00000277796;r=CHR_HSCHR17_10_CTG4:36194906-36196795;t=ENST00000612067 and http://www.ensembl.org/Homo_sapiens/Gene/Summary?g=ENSG00000276085;r=17:36194869-36196758). My question is, may I use the CCL3L3 count number as the CCL3L1 gene counts to calculate the IPS?
Additionally, the biomaRt package for R and the annotated gtf file from TCGA (https://api.gdc.cancer.gov/data/fe1750e4-fc2d-4a2c-ba21-5fc969a24f27) use the CAVIN2 as the alias for SDPR gene. So, this is a source of error running IPS with TCGA hg38 data.
Looking forward to hear from you
Cheers!
daniel