The opencadd.structure.pocket module

Let’s walk through the functionalities offered in the opencadd.structure.pocket module.

[1]:

%load_ext autoreload
%autoreload 2

[2]:

from pathlib import Path

import pandas as pd

from opencadd.databases.klifs import setup_remote, setup_local
from opencadd.structure.pocket import Pocket, PocketKlifs, PocketViewer

[3]:

HERE = Path(_dh[-1])
KLIFS_DOWNLOAD_PATH = HERE / "../../opencadd/tests/data/klifs"

Get example protein structure and pocket residues

[4]:

from opencadd.databases.klifs import setup_remote
KLIFS_REMOTE = setup_remote()

Fetch protein structure file content

First of all, we download structural data for an example protein kinase from the KLIFS database (some pocket residues are missing):

https://klifs.net/details.php?structure_id=12347

[5]:

structure_klifs_id = 12347

[6]:

text = KLIFS_REMOTE.coordinates.to_text(structure_klifs_id, extension="pdb")

Fetch pocket residues (or use your own pocket residues)

[7]:

pocket_residues = KLIFS_REMOTE.pockets.by_structure_klifs_id(structure_klifs_id)
pocket_residues

[7]:
residue.klifs_id residue.id residue.klifs_region_id residue.klifs_region residue.klifs_color
0 1 461 I.1 I khaki
1 2 462 I.2 I khaki
2 3 463 I.3 I khaki
3 4 _ g.l.4 g.l green
4 5 _ g.l.5 g.l green
... ... ... ... ... ...
80 81 594 xDFG.81 xDFG cornflowerblue
81 82 595 xDFG.82 xDFG cornflowerblue
82 83 _ xDFG.83 xDFG cornflowerblue
83 84 _ a.l.84 a.l cornflowerblue
84 85 _ a.l.85 a.l cornflowerblue

85 rows × 5 columns

The variables pocket_residue_ids and pocket_residue_ixs contain the list of residue PDB IDs and residue indices (derived from KLIFS sequence- and structure-based alignment). We will need this pocket information in the next step where we want to set up a pocket from opencadd’s Pocket class.

[8]:

pocket_residue_ids = pocket_residues["residue.id"].to_list()
print("Pocket residue PDB IDs:")
print(*pocket_residue_ids)
pocket_residue_ixs = pocket_residues["residue.klifs_id"].to_list()
print("Pocket residue (KLIFS) indices:")
print(*pocket_residue_ixs)

Pocket residue PDB IDs:
461 462 463 _ _ _ _ 468 469 470 471 472 473 480 481 482 483 484 485 497 498 499 500 501 502 503 504 505 506 507 508 509 511 512 513 514 515 516 517 518 519 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 592 593 594 595 _ _ _
Pocket residue (KLIFS) indices:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85

Pocket (Pocket class)

The Pocket class currently holds the following attributes/properties:

  • name: Protein/pocket name

  • filepath: Path to file with structural protein data

  • centroid: Centroids of all pocket residues’ CA atoms

  • subpockets: Subpockets defined based on a set of anchor residues each

  • regions: User-defined regions that are of importance for the protein/pocket

  • anchor_residues: Anchor residues to define one or more subpockets

Initialize pocket

We initialize the pocket with the following parameters:

  • Protein structure data

  • Protein/pocket name

  • Pocket residues PDB IDs

  • Pocket residue indices (optionally), e.g. for the pocket alignment IDs

[9]:

pocket = Pocket.from_text(
    text,
    "pdb",
    pocket_residue_ids,
    pocket_residue_ixs,
    name=structure_klifs_id
)

Let’s take a look at key Pocket class attributes/properties after initialization.

Pocket residues

All residue PDB IDs that cannot be cast to an integer are set to None.

[10]:

pocket.residues

[10]:
residue.id residue.ix
0 461 1
1 462 2
2 463 3
3 <NA> 4
4 <NA> 5
... ... ...
80 594 81
81 595 82
82 <NA> 83
83 <NA> 84
84 <NA> 85

85 rows × 2 columns

[11]:

print(*pocket._residue_ids)

461 462 463 None None None None 468 469 470 471 472 473 480 481 482 483 484 485 497 498 499 500 501 502 503 504 505 506 507 508 509 511 512 513 514 515 516 517 518 519 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 592 593 594 595 None None None

[12]:

print(*pocket._residue_ixs)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85

Pocket centroids

[13]:

pocket.center

[13]:

array([ 0.8315384, 21.615948 , 36.450153 ], dtype=float32)

Pocket data (atoms as DataFrame)

[14]:

pocket.data

[14]:
atom.id atom.name atom.x atom.y atom.z residue.id residue.name
73 74 N 9.014 18.239000 51.860001 461 GLN
74 75 CA 8.811 16.811001 51.655998 461 GLN
75 76 C 8.559 16.492001 50.186001 461 GLN
76 77 O 7.851 17.224001 49.491001 461 GLN
77 78 CB 7.630 16.325001 52.500999 461 GLN
... ... ... ... ... ... ... ...
1015 1016 CD1 -0.042 21.966000 31.419001 595 PHE
1016 1017 CD2 1.547 22.907000 29.912001 595 PHE
1017 1018 CE1 -0.337 23.228001 31.899000 595 PHE
1018 1019 CE2 1.255 24.172001 30.389000 595 PHE
1019 1020 CZ 0.311 24.332001 31.384001 595 PHE

577 rows × 7 columns

Add subpockets

Next, we can add subpockets one-by-one to the pocket. For each subpocket we define the following:

  • Subpocket name

  • Residue PDB IDs OR residue indices (e.g. alignment indices) of all anchor residues, i.e. the residues determining the subpocket center (centroid of all anchor residues’ CA atoms)

  • Subpocket color

The class method add_subpocket uses the Subpocket class to set up subpockets.

[15]:

pocket.add_subpocket("hinge_region", anchor_residue_ixs=[16, 47, 80], color="magenta")
pocket.add_subpocket("dfg_region", anchor_residue_ixs=[19, 24, 81], color="cornflowerblue")
pocket.add_subpocket("front_pocket", anchor_residue_ixs=[10, 48, 72], color="cyan")

Using the Pocket’s property subpockets, we get an overview of all specified subpockets.

[16]:

pocket.subpockets

[16]:
subpocket.name subpocket.color subpocket.center
0 hinge_region magenta [2.4566665, 22.592667, 41.674]
1 dfg_region cornflowerblue [8.460667, 20.395666, 33.809666]
2 front_pocket cyan [0.6393334, 16.937, 39.594666]

Using the Pocket‘s property anchor_residues, we get an overview of all subpockets’ anchor residues.

[17]:

pocket.anchor_residues

[17]:
subpocket.name anchor_residue.color anchor_residue.id anchor_residue.id_alternative anchor_residue.ix anchor_residue.center
0 hinge_region magenta 482 None 16 [8.327, 22.785, 43.461]
1 hinge_region magenta 531 None 47 [-0.001, 24.108, 46.55]
2 hinge_region magenta 593 None 80 [-0.956, 20.885, 35.011]
3 dfg_region cornflowerblue 485 None 19 [15.03, 17.46, 38.074]
4 dfg_region cornflowerblue 501 None 24 [8.524, 25.293, 29.217]
5 dfg_region cornflowerblue 594 None 81 [1.828, 18.434, 34.138]
6 front_pocket cyan 470 None 10 [11.026, 15.031, 41.426]
7 front_pocket cyan 532 None 48 [-3.527, 22.678, 46.33]
8 front_pocket cyan 578 None 72 [-5.581, 13.102, 31.028]

Subpockets are calculated based on so class anchor residues, defined each in an AnchorResidue class. Subpocket centers are the centroids of all anchor residues’ centers (i.e. normally the CA atoms).

  • If the anchor residue’s CA atom is available in the input structure is available, its coordinates are defined as the anchor residues center.

  • If the anchor residue’s CA atom is missing in a structure, alternative anchors are chosen if possible: If the residue CA atoms before and after the input anchor residue are available, their CA atoms’ centroid is chosen.

  • If only one of the neighboring residues’ CA atoms is available, that single CA atoms is chosen.

  • If none of the anchor residue’s and neighboring residues’ CA atoms is available, no anchor residue center is defined.

The determination of anchor residues depends on the CA atom availablity of the user-defined anchor residue as well as the residue before and after.

Add regions

The Pocket class also allows to specify pocket regions, normally used to store key regions, such as the hinge region or the catalytic loop in kinases. This information can be used for pocket visualization.

The class method add_regions uses the Regions class to set up regions.

[18]:

pocket.add_region("hinge", residue_ixs=[46, 47, 48], color="magenta")
pocket.add_region("linker", residue_ixs=[49, 50, 51, 52], color="cyan")
pocket.add_region("xDFG", residue_ixs=[80, 81, 82, 83], color="cornflowerblue")

[19]:

pocket.regions

[19]:
region.name region.color residue.id residue.ix
0 hinge magenta 530 46
1 hinge magenta 531 47
2 hinge magenta 532 48
3 linker cyan 533 49
4 linker cyan 534 50
5 linker cyan 535 51
6 linker cyan 536 52
7 xDFG cornflowerblue 593 80
8 xDFG cornflowerblue 594 81
9 xDFG cornflowerblue 595 82
10 xDFG cornflowerblue <NA> 83

Visualize pocket

Besides the pocket, we also want to visualize the co-crystallized ligand (if any), so let’s fetch the ligand Expo ID.

[20]:

ligand_expo_id = KLIFS_REMOTE.structures.by_structure_klifs_id(structure_klifs_id)["ligand.expo_id"][0]

[21]:

viewer = PocketViewer()
viewer.add_pocket(pocket, ligand_expo_id=ligand_expo_id)

[22]:

viewer.viewer.render_image(trim=True, factor=2, transparent=True),

[22]:

(Image(value=b'', width='99%'),)

[23]:

# Static output
viewer.viewer._display_image()

[23]:
../_images/tutorials_structure_pocket_41_0.png

KLIFS pocket (PocketKlifs class)

The PocketKlifs class is a child of the Pocket class, setting the kinase pocket regions as defined by KLIFS.

ba

Figure 1: Kinase pocket regions as defined by KLIFS (taken from the KLIFS publication)

Define subpockets (name and color) based on user-defined KLIFS residue IDs.

[24]:

subpockets = {
    "anchor_residue.klifs_ids": [[16, 47, 80], [19, 24, 81], [10, 48, 72]],
    "subpocket.name": ["hinge_region", "dfg_region", "front_pocket"],
    "subpocket.color": ["magenta", "cornflowerblue", "cyan"]
}
subpockets = pd.DataFrame(subpockets)
subpockets

[24]:
anchor_residue.klifs_ids subpocket.name subpocket.color
0 [16, 47, 80] hinge_region magenta
1 [19, 24, 81] dfg_region cornflowerblue
2 [10, 48, 72] front_pocket cyan

Initialize KLIFS pocket

… from remote KLIFS session

[25]:

kinase_pocket = PocketKlifs.from_structure_klifs_id(structure_klifs_id, subpockets)

This will internally, initiate a remote KLIFS session to fetch the relevant data. If you have a remote KLIFS session already initialized, you can also use it directly.

[26]:

kinase_pocket = PocketKlifs.from_structure_klifs_id(
    structure_klifs_id,
    subpockets,
    klifs_session=KLIFS_REMOTE
)

… from local KLIFS session

Use an example local KLIFS dataset.

[27]:

KLIFS_LOCAL = setup_local(KLIFS_DOWNLOAD_PATH)

… based on mol2 files

[28]:

kinase_pocket = PocketKlifs.from_structure_klifs_id(
    structure_klifs_id,
    subpockets,
    extension="mol2",
    klifs_session=KLIFS_LOCAL
)

Suspicious residue ID: _0 (from QH1_0)

… based on pdb files

[29]:

kinase_pocket = PocketKlifs.from_structure_klifs_id(
    structure_klifs_id,
    subpockets,
    extension="pdb",
    klifs_session=KLIFS_LOCAL
)

Visualize pocket with all KLIFS-defined regions

[30]:

viewer = PocketViewer()
viewer.add_pocket(kinase_pocket)

[31]:

viewer.viewer.render_image(trim=True, factor=2, transparent=True),

[31]:

(Image(value=b'', width='99%'),)

[32]:

# Static output
viewer.viewer._display_image()

[32]:
../_images/tutorials_structure_pocket_60_0.png