A USER ERROR has occurred: Positive training model failed to converge.

Question

[DareDevil]

I performed variant reclibration on 15 vcf files. But 1 vcf file encountered user error as follows:

A USER ERROR has occurred: Positive training model failed to converge.  One or more annotations (usually MQ) may have insufficient variance.  Please consider lowering the maximum number of Gaussians allowed for use in the model (via --max-gaussians 4, for example).

I know the solution is to reduce the --max-gaussians parameter. If I reduce it only for failed file will it affect my analysis? or should I keep the --max-gaussians same for all the 15 vcf files?

[DareDevil]

--max-gaussians default value is 8. I have changed to 6 for that particular sample

Answer 1

[User]

Have you received the wariming "Model could not pre-compute denominators. Denominator for gaussian evaluation cannot be computed." ?