Binder design workflow:
Initial length screening: First, use Bagel to perform exhaustive searches across different sequence lengths (80, 120, 160, and 200 amino acids) to determine the most suitable starting length range.
Region-specific design: Based on the selected optimal length, designs are carried out for different functional regions of RBX1:
Backbone selection and sequence design: Select top-ranking backbone structures from Bagel outputs and perform sequence design and optimization using SoluMPNN.
Rule-based filtering and iterative optimization: Apply heuristic rules summarized from the Bits to Binders competition for initial filtering. Select sequences with top global scores and re-input them into Bagel for further optimization.
Orthogonal evaluation and final selection: Use AF2 for orthogonal validation of high-ranking designs generated by Bagel, and select the final candidate binder sequences.
id: gentle-kiwi-pearl

RBX1
0.54
42.08
--
14.0 kDa
120
id: crimson-quail-ivy

RBX1
0.00
53.04
--
15.5 kDa
132
id: amber-mole-onyx

RBX1
0.46
38.23
--
18.5 kDa
160
id: gentle-zebra-iron

RBX1
0.10
44.32
--
15.1 kDa
132
id: frozen-bee-frost

RBX1
0.01
43.08
--
15.6 kDa
132
id: small-ram-bronze

RBX1
0.01
56.63
--
18.0 kDa
160
id: azure-fox-ash

RBX1
0.48
54.23
--
18.1 kDa
160
id: shy-zebra-leaf

RBX1
0.39
43.03
--
17.6 kDa
160
id: shy-tiger-cloud

RBX1
0.30
42.12
--
17.5 kDa
160
id: silent-goat-lotus

RBX1
0.72
44.55
--
14.8 kDa
123
id: shy-jaguar-lava

RBX1
0.45
41.26
--
14.3 kDa
123
id: quiet-owl-jade

RBX1
0.01
36.89
--
17.6 kDa
160
id: swift-tiger-pearl

RBX1
0.24
44.25
--
14.3 kDa
123
id: jade-bison-flint

RBX1
0.14
36.50
--
16.1 kDa
132
id: azure-raven-quartz

RBX1
0.00
28.63
--
17.9 kDa
160
id: swift-dove-iron

RBX1
0.35
30.98
--
14.2 kDa
120
id: gentle-vole-crystal

RBX1
0.30
35.72
--
18.9 kDa
160
id: quiet-raven-lava

RBX1
0.05
48.50
--
14.6 kDa
123