Versions

Change¶

Version	Date
`2022-2025`
`0.0.1`

Before `0.0.1`¶

Before the release 0.0.1, Drutai was used to infer a drug-target interaction per instance only. This process is simplified as follows.

Small molecule¶

Docetaxel (CID: 148124)

CC1=C2[C@H](C(=O)[C@@]3([C@H](C[C@@H]4[C@]([C@H]3[C@@H]([C@@](C2(C)C)(C[C@@H]1OC(=O)[C@@H]([C@H](C5=CC=CC=C5)NC(=O)OC(C)(C)C)O)O)OC(=O)C6=CC=CC=C6)(CO4)OC(=O)C)O)C)O

Target¶

A0A0H2UXE9

>A0A0H2UXE9
MDFPQQLEACVKQANQALSRFIAPLPFQNTPVVETMQYGALLGGKRLRPFLVYATGHMFGVSTNTLDAPAAAVECIHAYSLIHDDLPAMDDDDLRRGLPTCHVKFGEANAILAGDALQTLAFSILSDANMPEVSDRDRISMISELASASGIAGMCGGQALDLDAEGKHVPLDALERIHRHKTGALIRAAVRLGALSAGDKGRRALPVLDKYAESIGLAFQVQDDILDVVGDTATLGKRQGADQQLGKSTYPALLGLEQARKKARDLIDDARQALKQLAEQSLDTSALEALADYIIQRNK

Inference¶

drutai -m LSTMCNN -d drutai/data/example/148124.txt -t drutai/data/example/A0A0H2UXE9.fasta -mf /the/path/you/prefer/model/lstmcnn -o ./out.drutai

Tensorflow¶

Major change in deep learing libraries

In Keras 3, the way to call models becomes completely different from that in Keras 2.

Table 1:The way to use models.

Item	TF Keras 2	TF Keras 3
model format	`.h5` or TF `SavedModel`	`.keras` (new) + `.h5` (legacy)
`subclass` Models	supported via `SavedModel`	`TFSMLayer` for loading

Comparison¶

Our experience suggests that changes in Python versions have a smaller impact compared to changes in TensorFlow versions.

Table 2:The way to use models.

Python	Tensorflow verisons	Supported versions
`3.11`	🌟`2.14`🌟, `2.15`, `2.16`	TF Keras 2, TF Keras 3
`3.12`	`2.17`, `2.18`, `2.19`	only TF Keras 3