Waiting for the training script

[yunxiangz]

Thanks for providing the implementation of such wonderful work!
Can't wait to see the full training scripts as other Forkers do!

[luwei0917]

Thanks for your interest. we plan to release the training code in later August when the paper is accepted.

[yunxiangz]

Hi Wei,

Does the current version support the use of the trained model in a new protein's docking process, instead of the WDR domain of LRRK2 protein? I saw the coordinates of the pocket center are needed in the high_throughput_virtual_screening_LRRK2_WDR.ipynb, but didn't find the function of generating such coordinates of the pockets in of new protein.

Cheers.

[luwei0917]

Yes. You can construct the "info" pandas DataFrame yourself. The "pocket center" column could be manually changed.
Hope this help.

[PhungVanDuy]

@luwei0917 Do you have any updates for the training script?

[luwei0917]

Thanks for your interest.

In early commit (18314e9),
you could find an exmaple "examples/prediction_example_using_PDB_6hd6.ipynb" to construct
the dataset, and the main.py script to train. "python main.py -d 0 -m 0
--batch_size 5 --label baseline --addNoise 5
--use_equivalent_native_y_mask" is the script I used to train the model.

[PhungVanDuy]

Great! Thank you so much.

Thanks for your interest.

In early commit (18314e9), you could find an exmaple "examples/prediction_example_using_PDB_6hd6.ipynb" to construct the dataset, and the main.py script to train. "python main.py -d 0 -m 0 --batch_size 5 --label baseline --addNoise 5 --use_equivalent_native_y_mask" is the script I used to train the model.

That's great! Thank you so much.

[PhungVanDuy]

@luwei0917 Do you have any plan to publish the fully training dataset?

[luwei0917]

It has been uploaded.

[Yueming-Yin]

Hi Wei,

I ran the main.py as your script and got the following error:

File "TankBind-main\tankbind\data.py", line 164, in get_data
new_dataset.data = new_dataset.data.query("c_length < 100 and native_num_contact > 5").reset_index(drop=True)
AttributeError: 'NoneType' object has no attribute 'query'

Maybe the training data under "./apr22_pdbbind_gvp_pocket_radius20" is absent?

Thanks for your interest.

In early commit (18314e9), you could find an example "examples/prediction_example_using_PDB_6hd6.ipynb" to construct the dataset, and the main.py script to train. "python main.py -d 0 -m 0 --batch_size 5 --label baseline --addNoise 5 --use_equivalent_native_y_mask" is the script I used to train the model.

[theangle134]

Hi Wei,

I ran the main.py as your script and got the following error:

File "TankBind-main\tankbind\data.py", line 164, in get_data new_dataset.data = new_dataset.data.query("c_length < 100 and native_num_contact > 5").reset_index(drop=True) AttributeError: 'NoneType' object has no attribute 'query'

Maybe the training data under "./apr22_pdbbind_gvp_pocket_radius20" is absent?

hello,
I have the same mistake as you. Have you solved this problem? If so, can you help me

[Levi-ZJY]

Hi Wei,

I ran the main.py as your script and got the following error:

File "TankBind-main\tankbind\data.py", line 164, in get_data new_dataset.data = new_dataset.data.query("c_length < 100 and native_num_contact > 5").reset_index(drop=True) AttributeError: 'NoneType' object has no attribute 'query'

Maybe the training data under "./apr22_pdbbind_gvp_pocket_radius20" is absent?

Thanks for your interest.
In early commit (18314e9), you could find an example "examples/prediction_example_using_PDB_6hd6.ipynb" to construct the dataset, and the main.py script to train. "python main.py -d 0 -m 0 --batch_size 5 --label baseline --addNoise 5 --use_equivalent_native_y_mask" is the script I used to train the model.

hello,
I also have the same mistake... Have you solved it now?

[browna1]

Hi Wei,

I ran the main.py as your script and got the following error:

File "TankBind-main\tankbind\data.py", line 164, in get_data new_dataset.data = new_dataset.data.query("c_length < 100 and native_num_contact > 5").reset_index(drop=True) AttributeError: 'NoneType' object has no attribute 'query'

Maybe the training data under "./apr22_pdbbind_gvp_pocket_radius20" is absent?

Thanks for your interest.
In early commit (18314e9), you could find an example "examples/prediction_example_using_PDB_6hd6.ipynb" to construct the dataset, and the main.py script to train. "python main.py -d 0 -m 0 --batch_size 5 --label baseline --addNoise 5 --use_equivalent_native_y_mask" is the script I used to train the model.

hello!!
I also have this mistake, Have you solved it now?

[mrqfliu]

Hi Wei,

我将 main.py 作为您的脚本运行，并收到以下错误：

文件“TankBind-main\tankbind\data.py”，第 164 行，get_data new_dataset.data = new_dataset.data.query（“c_length < 100 和 native_num_contact > 5”）.reset_index（drop=True） AttributeError：“NoneType”对象没有属性“query”

也许“./apr22_pdbbind_gvp_pocket_radius20”下的训练数据不存在？

Thanks for your interest.
在早期提交 （18314e9） 中，您可以找到一个示例“examples/prediction_example_using_PDB_6hd6.ipynb”来构建数据集，以及要训练 main.py 脚本。“python main.py -d 0 -m 0 --batch_size 5 --label baseline --addNoise 5 --use_equivalent_native_y_mask”是我用来训练模型的脚本。

hello!!
I also have this mistake, Have you solved it now?