Question about the time it takes for VC analysis about deepvariant HOT 5 CLOSED

@bkurtoglu , can you please let us know what datatype you are using? The current model only works with ONT R10.4 simplex and duplex data. Can you provide some more details about your data and run:

1. What chemistry is your ONT data?
2. What basecaller did you use?
3. How many cpus are you using to run make_examples?

from deepvariant.

1. The type of the data is ONT_R9 simplex.
2. The basecaller is guppy.
3. Threads are set to 48.
   I also provide the code as below. Thank you for your help.
   BIN_VERSION="1.6.1" sudo apt -y update sudo apt-get -y install docker.io sudo docker pull google/deepvariant:"${BIN_VERSION}" INPUT_DIR="/home/user/Masaüstü/BEYZA/input" OUTPUT_DIR="/home/user/Masaüstü/BEYZA/output" THREADS=48 MODEL_NAME="ONT_R104" sudo docker run -v "/home/massive/Desktop/beyza/input":"/input" -v "/home/massive/Desktop/beyza/OUTPUTDEEP":"/output" google/deepvariant:"${BIN_VERSION}" /opt/deepvariant/bin/run_deepvariant --model_type=ONT_R104 --ref="/input/GCA_000001405.15_GRCh38_no_alt_analysis_set.fna" --reads="/input/NA12878-minion-ul_GRCh38.bam" –output_vcf="/output/deep.vcf" --threads=48

from deepvariant.

@bkurtoglu ,

ONT_R9 has a very high error rate, DeepVariant isn't able to work on it directly. You can try our previous solution PEPPER for R9 data. It's also mentioned on our README page:

• Oxford Nanopore R9.4.1 data by using PEPPER-DeepVariant.

So it's unlikely you will be successful to run DeepVariant R10.4 model on R9 data.

from deepvariant.

@kishwarshafin
Thanks a lot for clarifying that, we were making a benchmark study on variant callers to optimize our pipeline. We also used PEPPER-DeepVariant and it worked with high precision. Again, thanks a lot! It was a helpful communication.

from deepvariant.

Great, thanks, I will close the issue.

from deepvariant.