# Number of targets and antitargets to select
protein_selection:
  small_molecule:
    target: "Q9UQM7" # if multiple targets, they should be comma separated
    clip_interval: [[340, 475]] # interval to clip sequence (0-based, exclude aas outside of this interval). If None, no clipping is done. One value for each target.
  nanobody:
    target: "P01584" # if multiple targets, they should be comma separated
    clip_interval: [[116, 269]] # interval to clip sequence (0-based, exclude aas outside of this interval). If None, no clipping is done. One value for each target.

# Competition parameters
competition:
  # No submission blocks
  no_submission_blocks: 10
  nanobody_weight: 0.6 # proportion of incentive to give to nanobody winner

# Bounty payout settings
payout:
  enabled: true
  override_uid: 61 # compound payout UID

# Molecular property requirements
molecule_requirements:
  num_molecules: 1
  min_heavy_atoms: 10
  min_rotatable_bonds: 1
  max_rotatable_bonds: 10
  min_entropy: 0.1 # minimum entropy required when num_molecules > 1
  banned_atom_types: ["Se", "Na", "Fe", "Zn"]

# Reaction filtering configuration
reaction_filtering:
  random_valid_reaction: true  # true = one random reaction per epoch, false = only allowed reactions below are allowed
  allowed_reactions: ["rxn:1", "rxn:2"] # allowed reactions (only used if random_valid_reaction is false)

# Boltz2 configuration
boltz2_config:
  boltz_metric: ["affinity_probability_binary", "affinity_pred_value"] # metric to use for scoring: ["affinity_pred_value", "affinity_probability_binary", "confidence_score", ...]
  combination_strategy: "heavy_atom_normalization" # "heavy_atom_normalization" or "average"
  boltz_mode: max # "max" or "min"

# Nanobody requirements
nanobody_requirements:
  num_sequences: 1
  min_sequence_length: 90
  max_sequence_length: 150
  min_cysteines: 1 # minimum number of cysteines required
  cys_pair_min_separation: 35 # minimum separation between cysteine pairs, only valid when min_cysteines > 1
  cys_pair_max_separation: 80 # maximum separation between cysteine pairs, only valid when min_cysteines > 1
  max_homopolymer_run: 6 # maximum length of a homopolymer run (e.g. AAAAAA)
  max_di_repeat_pairs: 4 # maximum number of di-repeat pairs (e.g."GSGSGSGSGS" has 5 pairs -> reject)
  reject_signal_peptides: true # reject sequences that look like signal peptides (strong N term hydrophobicity)
  sp_window: 12
  sp_hydro_min_in_window: 8 # >=8 hydrophobic in a 12-aa window
  sp_scan_prefix: 30 # length of N term sequence to scan for signal peptide
  min_nativeness_score: 0.45
  min_human_framework_score: 0.75
  max_similarity_score: 0.95 # maximum similarity score to top sequences for that target (calculated by metanano.search.search_engine)
  num_top_sequences: 50 # number of top sequences to consider for similarity check
  enforce_vhh_hallmarks: true # enforce VHH FR2 region hallmark residues (E/Q at pos 49, R at pos 50)

boltzgen_config:
  boltzgen_rank_mode: min # "max" or "min"
  boltzgen_rank_by: rank_sum # "rank_sum" or "worst_rank"