Hello everybody!

Several gnomAD posts (v.2.1 Exome) state that the gnomAD team applied a random forest for variant filtering. This filter tells us basically if a variant passes or fails depending on if the random forest value is >= 0.1 or not.

When checking the gnomAD database online, I came across several variants which have a RF value below 0.1 (e.g. 14:107049035:G:C), but still pass the filter and others with a value higher than 0.1, but failing the filter.

Any idea, what could be the problem here?

Best, Andreas