orca_models module¶

This module contains the class definition of all Orca models. For usage of the models, see the orca_predict module.

class orca_models.H1esc[source]¶

Bases: torch.nn.modules.module.Module

Orca H1-ESC model (1-32Mb)

net0¶

The first section of the multi-resolution encoder (bp resolution to 4kb resolution).

Type: nn.DataParallel(Encoder)

net¶

The second section of the multi-resolution encoder (4kb resolution to 128kb resolution).

Type: nn.DataParallel(Encoder2)

denets¶

Decoders at each level, which are stored in a dictionary with an integer as key.

Type: dict(int: nn.DataParallel(Decoder))

normmats¶

The distance-based background matrices with expected log fold over background values at each level.

Type: dict(int: numpy.ndarray)

epss¶

The minimum background value at each level. Used for stablizing the log fold computation by adding to both the nominator and the denominator.

Type: dict(int: float)

class orca_models.H1esc_1M[source]¶

Bases: torch.nn.modules.module.Module

Orca H1-ESC model (1Mb)

net¶

Integrated Encoder and Decoder for 1Mb model.

Type: nn.DataParallel(Net)

normmats¶

The distance-based background matrices with expected log fold over background values at each level.

Type: dict(int: numpy.ndarray)

epss¶

The minimum background value at each level. Used for stablizing the log fold computation by adding to both the nominator and the denominator.

Type: dict(int: float)

forward(x)[source]¶

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class orca_models.H1esc_256M[source]¶

Bases: torch.nn.modules.module.Module

Orca H1-ESC model (32-256Mb)

net0¶

The first section of the multi-resolution encoder (bp resolution to 4kb resolution).

Type: nn.DataParallel(Encoder)

net1¶

The second section of the multi-resolution encoder (4kb resolution to 128kb resolution).

Type: nn.DataParallel(Encoder2)

net¶

The third section of the multi-resolution encoder (128kb resolution to 1024kb resolution).

Type: nn.DataParallel(Encoder3)

denets¶

Decoders at each level, which are stored in a dictionary with an integer as key.

Type: dict(int: nn.DataParallel(Decoder))

normmats¶

The distance-based background matrices with expected log fold over background values at each level.

Type: dict(int: numpy.ndarray)

epss¶

The minimum background value at each level. Used for stablizing the log fold computation by adding to both the nominator and the denominator.

Type: dict(int: float)

class orca_models.HCTnoc[source]¶

Bases: torch.nn.modules.module.Module

Orca cohesin-depleted HCT116 model (1-32Mb)

net0¶

The first section of the multi-resolution encoder (bp resolution to 4kb resolution).

Type: nn.DataParallel(Encoder)

net¶

The second section of the multi-resolution encoder (4kb resolution to 128kb resolution).

Type: nn.DataParallel(Encoder2)

denets¶

Decoders at each level, which are stored in a dictionary with an integer as key.

Type: dict(int: nn.DataParallel(Decoder))

normmats¶

The distance-based background matrices with expected log fold over background values at each level.

Type: dict(int: numpy.ndarray)

epss¶

The minimum background value at each level. Used for stablizing the log fold computation by adding to both the nominator and the denominator.

Type: dict(int: float)

class orca_models.Hff[source]¶

Bases: torch.nn.modules.module.Module

Orca HFF model (1-32Mb)

net0¶

The first section of the multi-resolution encoder (bp resolution to 4kb resolution).

Type: nn.DataParallel(Encoder)

net¶

The second section of the multi-resolution encoder (4kb resolution to 128kb resolution).

Type: nn.DataParallel(Encoder2)

denets¶

Decoders at each level, which are stored in a dictionary with an integer as key.

Type: dict(int: nn.DataParallel(Decoder))

normmats¶

The distance-based background matrices with expected log fold over background values at each level.

Type: dict(int: numpy.ndarray)

epss¶

The minimum background value at each level. Used for stablizing the log fold computation by adding to both the nominator and the denominator.

Type: dict(int: float)

class orca_models.Hff_1M[source]¶

Bases: torch.nn.modules.module.Module

Orca HFF model (1Mb)

net¶

Integrated Encoder and Decoder for 1Mb model.

Type: nn.DataParallel(Net)

normmats¶

The distance-based background matrices with expected log fold over background values at each level.

Type: dict(int: numpy.ndarray)

epss¶

The minimum background value at each level. Used for stablizing the log fold computation by adding to both the nominator and the denominator.

Type: dict(int: float)

forward(x)[source]¶

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class orca_models.Hff_256M[source]¶

Bases: torch.nn.modules.module.Module

Orca HFF model (32-256Mb)

net0¶

The first section of the multi-resolution encoder (bp resolution to 4kb resolution).

Type: nn.DataParallel(Encoder)

net1¶

The second section of the multi-resolution encoder (4kb resolution to 128kb resolution).

Type: nn.DataParallel(Encoder2)

net¶

The third section of the multi-resolution encoder (128kb resolution to 1024kb resolution).

Type: nn.DataParallel(Encoder3)

denets¶

Decoders at each level, which are stored in a dictionary with an integer as key.

Type: dict(int: nn.DataParallel(Decoder))

normmats¶

The distance-based background matrices with expected log fold over background values at each level.

Type: dict(int: numpy.ndarray)

epss¶

The minimum background value at each level. Used for stablizing the log fold computation by adding to both the nominator and the denominator.

Type: dict(int: float)