Table 1 Scientific tests for bio-macromolecular modeling, continuously running on our testing server framework.

From: Ensuring scientific reproducibility in bio-macromolecular modeling via extensive, automated benchmarks

Test suite

Tests

Refs.

Test author

Quality measures

Targets

nstruct

Runtime in CPUh

Antibodies

antibody_grafting

55

Jeliazko Jeliazkov

Fraction residues within rmsd to native

48

1

3

antibody_h3_modeling

56

Score vs. rmsd

6

500

3000

antibody_snugdock

57

I_sc vs. I_rmsd

6

500

3000

Carbohydrates

glycan_dock, (dock_glycans)*

58,59

Jason Labonte, Morgan Nance

I_sc vs. L_rmsd

6

1000

1100

glycan_structure_prediction

60

Jared Adolf-Bryfogle

Score vs. rmsd

4

500

950

Comparative modeling

RosettaCM

61

Jason Fell

GDT-MM

16

200

1800

Design

ddg_alanine_scan

62

Ajasja Ljubetič

R, MAE, fraction correctly classified

19: 381

1

3

Design

SEWING

63

Frank Teets

MotifScorer, InterModelMotifScorer

1

100

75

Design

enzyme_design

64

Rocco Moretti

Various sequence recoveries

50

1

50

Design

design_fast

65

Jack Maguire, Chris Bahl

Score vs. seqrec

48

100

2600

Design, interfaces

cofactor_binding_sites

66

Amanda Loshbaugh

rank top, position profile similarity

7

200

170

design, immune system

mhc_epitope_energy

67

Brahm Yachnin

Degree of de-immunization, among others

50

100

2000

docking

protein_protein_docking

68

Shourya SR Burman

I_sc vs. I_rmsd

10

5000

833

ensemble docking

69

Ameya Hamalkar

I_sc vs I_rmsd

3

5000

3000

FlexPepDock

FlexPepDock

70

Ziv Ben-Aharon

reweighted I_sc vs backbone I_rmsd

2

200

70

fragments

fragment_picking

71

Justyna Krys, Dominik Gront

rmsd

10

400

2000

fragments

make fragments pipeline

71

Daniel Farrell

Coverage, precision

65

1

3000

ligand docking

ligand_docking

50

Shannon Smith

Delta_Isc vs. ligand_rmsd

50

200

2000

ligand_scoring_ranking

50

Spearman and Pearson correlation coefficient

57: 285

1

2

loop modeling

loop_modeling_CCD

72

Phuong Tran, Shane Ó Conchúir

Score vs. loop_rmsd

7

500

500

loop_modeling_KIC

73

Score vs. loop_rmsd

7

500

620

loop_modeling_KIC_fragments

74

Score vs. loop_rmsd

7

500

760

loop_modeling_NGK

75

Score vs. loop_rmsd

7

500

570

membrane protein-energy function

mp_f19_energy_landscape#

37

Rituparna Samanta, Rebecca Alford

ddG, depth and title angle

4

1

10

mp_f19_decoy_discrimination

37

Score vs. rmsd, Wrms

4×100

1

2000

mp_f19_sequence_recovery

37

sequence recovery, Kullback-Leibler divergence

130

1

500

mp_f19_ddG_of_mutation

76

Pearson correlation coefficient

3

1

1

membrane proteins

mp_dock

77

Julia Koehler Leman, Rebecca Alford

I_sc vs. I_rmsd

10

1000

200

mp_domain_assembly

78

Score vs. rmsd

5

5000

700

mp_lipid_acc

79

Accuracy

223

1

2

mp_relax

77

Score vs. rmsd

4

100

40

mp_symdock

77

I_sc vs. rmsd

5

1000

140

PDB diagnostic

PDB_diagnostic

NA

Steven Lewis, William Hansen, Sergey Lyskov

Read-in error type

entire PDB

1

1000

peptide structure prediction

simple_cycpep_predict

48

Vikram K. Mulligan

Score vs. rmsd, PNear

1

~800,000

320

peptide_pnear_vs_ic50

51

IC50 vs. folding energy

7

80,000

400

refinement

relax_cartesian

32

Julia Koehler Leman

Score vs. rmsd

12

100

120

relax_fast

80

Score vs. rmsd

12

100

120

relax_fast_5iter

80

Score vs. rmsd

12

100

120

RNA

rna_denovo_favorites

81

Andy Watkins

Score vs. rmsd

12

200

120

stepwise_rna_favorites

82

Score vs. rmsd

12

200

240

RosettaNMR

abinitio_RosettaNMR_rdc

83

Georg Kuenze, Julia Koehler Leman

Score vs. rmsd

3

2000

170

abinitio_RosettaNMR_pcs

83

Score vs. rmsd

3

2000

1400

  1. The number of tests is constantly being expanded. The test suite is the overall application, the test is the specific test, implemented by the test author(s). The quality measures are evaluated to choose a pass/fail criterion. The targets are the number of different proteins (or biomolecules) tested on, nstruct is the number of models built for each target, and the runtime in CPU hours is the total runtime over all targets.
  2. *The dock_glycans test has been superceded by glycan_dock.
  3. #The mp_f19_energy_landscape test has been renamed to mp_f19_tilt_angle.