kipoi_veff.scores
Ref
Ref(self, rc_merging='mean')
Returns the predictions for the reference allele.
If the predictions were executed taking
the reverse-complement of the sequence into account then the returned value is averaged by
the function defined in rc_merging
. Allowed values for rc_merging
are: "min", "max", "mean", "median",
"absmax" or any callable that accepts/expects two arguments: my_func(fwd_pred, rc_pred)
.
Reverse-complement-averaging, where applicable, is performed after score calculation.
Alt
Alt(self, rc_merging='mean')
Returns the predictions for the alternative allele.
If the predictions were executed taking
the reverse-complement of the sequence into account then the returned value is averaged by
the function defined in rc_merging
. Allowed values for rc_merging
are: "min", "max", "mean", "median",
"absmax" or any callable that accepts/expects two arguments: my_func(fwd_pred, rc_pred)
.
Reverse-complement-averaging, where applicable, is performed after score calculation.
Diff
Diff(self, rc_merging='mean')
Returns the difference between predictions for the reference and alternative
sequences prediction difference: diff = p_alt - p_ref
If the predictions were executed taking
the reverse-complement of the sequence into account then the returned value is averaged by
the function defined in rc_merging
. Allowed values for rc_merging
are: "min", "max", "mean", "median",
"absmax" or any callable that accepts/expects two arguments: my_func(fwd_pred, rc_pred)
.
Reverse-complement-averaging, where applicable, is performed after score calculation.
LogitRef
LogitRef(self, rc_merging='mean')
Returns the predictions for the reference allele on the logit scale: np.log(p_alt / (1 - p_alt ))
.
If the predictions were executed taking
the reverse-complement of the sequence into account then the returned value is averaged by
the function defined in rc_merging
. Allowed values for rc_merging
are: "min", "max", "mean", "median",
"absmax" or any callable that accepts/expects two arguments: my_func(fwd_pred, rc_pred)
.
Reverse-complement-averaging, where applicable, is performed after score calculation.
LogitAlt
LogitAlt(self, rc_merging='mean')
Returns the predictions for the alternative allele on the logit scale: np.log(p_alt / (1 - p_alt ))
If the predictions were executed taking
the reverse-complement of the sequence into account then the returned value is averaged by
the function defined in rc_merging
. Allowed values for rc_merging
are: "min", "max", "mean", "median",
"absmax" or any callable that accepts/expects two arguments: my_func(fwd_pred, rc_pred)
.
Reverse-complement-averaging, where applicable, is performed after score calculation.
Logit
Logit(self, rc_merging='mean')
Returns the difference between predictions for the reference and alternative sequences on
the logit scale: logit_diff = log(p_alt / (1 - p_alt )) - log(p_ref / (1 - p_ref ))
If the predictions were executed taking
the reverse-complement of the sequence into account then the returned value is averaged by
the function defined in rc_merging
. Allowed values for rc_merging
are: "min", "max", "mean", "median",
"absmax" or any callable that accepts/expects two arguments: my_func(fwd_pred, rc_pred)
.
Reverse-complement-averaging, where applicable, is performed after score calculation.
DeepSEA_effect
DeepSEA_effect(self, rc_merging='mean')
Returns the score used by DeepSEA in order to calculate the e-value: abs(logit_diff) * abs(diff)
If the predictions were executed taking
the reverse-complement of the sequence into account then the returned value is averaged by
the function defined in rc_merging
. Allowed values for rc_merging
are: "min", "max", "mean", "median",
"absmax" or any callable that accepts/expects two arguments: my_func(fwd_pred, rc_pred)
.
Reverse-complement-averaging, where applicable, is performed after score calculation.