0 % of model's snps used for trans QTL

[elnetwas]

Hello MetaXcan,
I am attempting to use predict.py and receiving the INFO - 0 % of models' snps used message. My model in db format is similar to that of elastic net, obtained using the tensorQTL tool, where I perform a trans-mapping. I format these QTLs appropriately for elastic net with 57,997 variants, resulting in the following format:

Considering that the gene variable is predetermined and always set as ENSG0000269981, as it pertains to a single phenotype, and my imputed genotype (which is the same one used in tensorQTL to obtain the above-mentioned variants) is in the following format:

I apply the following script using Predict.py:
python3 $METAXCAN/Predict.py --model_db_path trans_associations_mucosa.db --model_db_snp_key varID --vcf_genotypes /mnt/lustre/scratch/nlsas/home/otras/fmx/fgs/carmen/telomeros/GENOTIPO/resultados/GENOTIPOS_IMPUTADOS_FILTRADO_FASTQTL.vcf.gz --vcf_mode genotyped --prediction_output trans_associations_mucosa_umbral_dos_predict.txt --prediction_summary_output trans_associations_mucosa_umbral_dos_summary.txt --verbosity 9 --throw --on_the_fly_mapping METADATA "{}{}{}_{}_b37"
However, I consistently get "0 % of models' snps used." I have several questions: Can predict.py be applied to trans-QTLs? Or is it possible that the low number of variants is causing the result of "0% of models' snps used"?

Thank you for your time!!