The Signalp v3 nn model will raise an error if an input sequence is too long.
It's unclear what the threshold is.
Maybe 5000AA -ish?
Currently we just ignore the error and live with it as missing data.
The more correct approach might be to truncate the proteins first, and run it through like that.
I don't think it's critical ATM.
For future perfectionism.
The Signalp v3 nn model will raise an error if an input sequence is too long. It's unclear what the threshold is. Maybe 5000AA -ish?
Currently we just ignore the error and live with it as missing data. The more correct approach might be to truncate the proteins first, and run it through like that.
I don't think it's critical ATM. For future perfectionism.