get prediction of my own test data

[Ankitapal1995]

I have few sequences I want to run MTB-CNN model and check the result. But I am confused about data preparation from fasta/fastq sequence files and run the scripts for drug resistance prediction. Could you please guide me?
Thanks in advance.

[aggreen]

Hello, the steps for going from fastq to drug resistance prediction will be as follows:

1. use a variant calling pipeline to produce .vcf files, relative to the reference H37Rv. Our study used isolates that were previously processed through the pipeline described in this manuscript: https://www.nature.com/articles/s41467-021-26248-1
2. Use the code in input_data/prep_fasta_files to convert the .vcf files into .fasta files aligned to the H37Rv reference
3. Align your .fasta files to the .fasta files previously used by the model, to preserve the frame of reference, as explained in the readme

At this point, the files should be ready for prediction!

[Ankitapal1995]

Thank you so much for guiding me.