config.yaml

# ============================================================
# CSF Autophagy/Lysosome Panel Discovery Pipeline Configuration
# ============================================================

project:
  name: "CSF_Autophagy_Lysosome_Panel"
  version: "1.0"
  random_seed: 42

paths:
  raw_dir: "../../raw"
  derived_dir: "../DerivedData"
  qc_dir: "../QC"
  output_dir: "../Outputs"
  log_dir: "../Log"

# -----------------------------------------------------------
# Dataset Registry
# -----------------------------------------------------------
datasets:
  D1:
    file: "The list of all identified proteins from lable-free MS of mouse CSF(Per).xlsx"
    sheet: "All identified proteins"
    format: "maxquant_fasta"
    species: "mouse"
    biofluid: "CSF"
    role: "mouse_csf"
    id_columns:
      protein_ids: "Protein IDs"
      majority_ids: "Majority protein IDs"
      fasta_headers: "Fasta headers"
    intensity_prefix: "LFQ intensity"
    intensity_scale: "raw"
    n_samples_expected: 12
    notes: "Presence-only for scoring (weak label-free); gene via GN= regex on Fasta headers"

  D2:
    file: "Pooled datasets blasting overview.xlsx"
    sheet: "mouse_csf_PXD053698"
    format: "maxquant_preparsed"
    species: "mouse"
    biofluid: "CSF"
    role: "mouse_csf"
    id_columns:
      gene_symbol: "gene_symbol"
      uniprot_id: "uniprot_id"
    intensity_prefix: "LFQ intensity"
    intensity_scale: "raw"
    n_samples_expected: 43
    meta_columns: ["gene_symbol", "uniprot_id", "protein_name", "num_samples_detected", "mean_LFQ_intensity"]

  D3:
    file: "Pooled datasets blasting overview.xlsx"
    sheet: "mouse_csf_PXD053568_5xFAD "
    format: "maxquant_preparsed"
    species: "mouse"
    biofluid: "CSF"
    role: "mouse_csf"
    id_columns:
      gene_symbol: "gene_symbol"
      uniprot_id: "uniprot_id"
    intensity_prefix: "LFQ intensity"
    intensity_scale: "raw"
    n_samples_expected: 10
    meta_columns: ["gene_symbol", "uniprot_id", "protein_name", "num_samples_detected", "mean_LFQ_intensity"]
    notes: "Sheet name has trailing space"

  D4:
    file: "Project PXD032782/APPPS1_BI_2_2Pep.xlsx"
    sheet: "Sheet1"
    format: "diann_excel"
    species: "mouse"
    biofluid: "CSF"
    role: "mouse_csf"
    id_columns:
      protein_group: "Protein.Group"
      protein_ids: "Protein.Ids"
      genes: "Genes"
    intensity_prefix: "LFQ intensity"
    intensity_scale: "raw"
    n_samples_expected: 32
    meta_columns: ["Protein.Group", "Protein.Ids", "Protein.Names", "Genes", "First.Protein.Description"]
    extra_column_prefixes_to_skip: ["Precursor", "Peptides"]
    notes: "Use 2-peptide file; has LFQ + Precursor + Peptides columns; use LFQ only"

  D5:
    file: "Project PXD065438/Output_CSF-mk_Trem2/CSF-mk_Trem2.pg_matrix.tsv"
    format: "diann_tsv"
    species: "mouse"
    biofluid: "CSF"
    role: "mouse_csf"
    id_columns:
      protein_group: "Protein.Group"
      genes: "Genes"
    intensity_scale: "raw"
    n_samples_expected: 15
    meta_columns: ["Protein.Group", "Protein.Names", "Genes", "First.Protein.Description"]
    notes: "No Protein.Ids column; sample cols are Windows file paths"

  D6:
    file: "Project PXD065438/Output_Brain-mk_Trem2/AGCH2501_brain.pg_matrix.tsv"
    format: "diann_tsv"
    species: "mouse"
    biofluid: "brain"
    role: "brain_plausibility"
    id_columns:
      protein_group: "Protein.Group"
      genes: "Genes"
    intensity_scale: "raw"
    n_samples_expected: 15
    meta_columns: ["Protein.Group", "Protein.Names", "Genes", "First.Protein.Description"]

  D7:
    file: "Project PXD052590/report.pg_matrix.tsv"
    format: "diann_tsv"
    species: "mouse"
    biofluid: "CSF"
    role: "mouse_csf"
    id_columns:
      protein_group: "Protein.Group"
      protein_ids: "Protein.Ids"
      genes: "Genes"
    intensity_scale: "raw"
    n_samples_expected: 20
    meta_columns: ["Protein.Group", "Protein.Ids", "Protein.Names", "Genes", "First.Protein.Description"]

  D8:
    file: "Project PXD048864/report.pg_matrix_CSF.tsv"
    format: "diann_tsv"
    species: "mouse"
    biofluid: "CSF"
    role: "mouse_csf"
    id_columns:
      protein_group: "Protein.Group"
      protein_ids: "Protein.Ids"
      genes: "Genes"
    intensity_scale: "raw"
    n_samples_expected: 49
    meta_columns: ["Protein.Group", "Protein.Ids", "Protein.Names", "Genes", "First.Protein.Description"]

  D9:
    file: "Pooled datasets blasting overview.xlsx"
    sheet: "mouse_isf_PDX048864"
    format: "diann_excel"
    species: "mouse"
    biofluid: "ISF"
    role: "isf_plausibility"
    id_columns:
      protein_group: "Protein.Group"
      protein_ids: "Protein.Ids"
      genes: "Genes"
    intensity_scale: "raw"
    n_samples_expected: 47
    meta_columns: ["Protein.Group", "Protein.Ids", "Protein.Names", "Genes", "First.Protein.Description"]
    notes: "2 completely empty sample columns to auto-detect and exclude"

  D10:
    file: "Pooled datasets blasting overview.xlsx"
    sheet: "human_csf_PXD053698"
    format: "maxquant_preparsed"
    species: "human"
    biofluid: "CSF"
    role: "human_csf_supporting"
    id_columns:
      gene_symbol: "gene_symbol"
      uniprot_id: "uniprot_id"
    intensity_prefix: "LFQ intensity"
    intensity_scale: "raw"
    n_samples_expected: 8
    meta_columns: ["gene_symbol", "uniprot_id", "protein_name", "num_samples_detected", "mean_LFQ_intensity"]

  D11:
    file: "Project Astral/Astral protome_discovery and repliction_DACboth013_49D_1_7_3_Sc01.tsv"
    format: "diann_astral"
    species: "human"
    biofluid: "CSF"
    role: "human_csf_primary"
    id_columns:
      protein: "protein"
      protein_group: "Protein.Group"
      genes: "Genes"
    intensity_scale: "log2"
    n_samples_expected: 2720
    meta_columns: ["protein", "Protein.Group", "Protein.Names", "Genes"]
    notes: "Primary human CSF evidence; log2 intensities; 3232 proteins"

  D12:
    file: "Project Astral/Astral proteome_repliction only_DACRP014_49D_1_7_2_Sc03.tsv"
    format: "diann_astral"
    species: "human"
    biofluid: "CSF"
    role: "human_csf_validation"
    id_columns:
      protein: "protein"
      protein_group: "Protein.Group"
      genes: "Genes"
    intensity_scale: "log2"
    n_samples_expected: 643
    precomputed_stats:
      median: "abundance_median_Astral"
      rank: "abundance_rank_Astral"
      observations: "observations_Astral"
      completeness: "data_completeness_Astral"
    meta_columns: ["protein", "Protein.Group", "Protein.Names", "Genes",
                    "abundance_median_Astral", "abundance_rank_Astral",
                    "observations_Astral", "data_completeness_Astral"]
    notes: "Validation only; never count as independent evidence"

  EV:
    file: "WT vs KO EV protein list from highest to lowest abundance.xlsx"
    format: "ev_rank"
    species: "human_cell_line"
    biofluid: "EV"
    role: "ev_annotation"
    sheets:
      raw: "Raw list Lovisa"  # NB: literal Excel sheet name in the source file
      precomputed: "330 after blast with full"
      overlap: "Overlap detection"
    notes: "Rank-ordered gene symbols; annotation only (weight 0.05)"

# -----------------------------------------------------------
# Reference Inputs
# -----------------------------------------------------------
references:
  R1:
    file: "Full list with autophagy proteins curated.xlsx"
    deduplicated_sheet: "Sheet1"
    full_sheet: "Full list"
    columns:
      gene_symbol: "Official Gene symbol"
      alias: "Alias (optional)"
      category: "Biological Function"
      description: "Description"
      refs: "Refs"

  R2:
    file: "Pooled datasets blasting overview.xlsx"
    sheet: "Overview"
    columns:
      antibody: "Antibody"
      gene_symbol: "GeneSymbol"
      abbreviation: "Abbreviation"

# -----------------------------------------------------------
# Evidence Score Weights
# -----------------------------------------------------------
scoring:
  weights:
    mouse_csf: 0.25
    human_csf: 0.30
    ev: 0.10
    brain_plausibility: 0.10
    autophagy_membership: 0.25

  penalties:
    ambiguous_group_unresolved: -0.10
    orthology_ambiguous_unresolved: -0.10
    likely_plasma_derived: -0.20

  human_csf_tier_scores:
    A: 1.0
    B: 0.7
    C: 0.3
    absent: 0.0

  d10_bonus: 0.1

  autophagy_category_modifiers:
    core_machinery: 1.0
    lysosomal: 1.0
    mitophagy: 0.85
    docking: 0.85
    upstream_regulators: 0.70
    default: 0.80

  mouse_csf_datasets: ["D1", "D2", "D3", "D4", "D5", "D7", "D8"]

# -----------------------------------------------------------
# Detectability Tier Definitions
# -----------------------------------------------------------
detectability:
  small_dataset_threshold: 20
  tiers:
    A:
      small:
        min_detected: 2
        above_floor: true
      large:
        min_fraction: 0.10
        above_floor: true
    B:
      small:
        min_detected: 1
        above_floor: true
      large:
        min_fraction: 0.02
        above_floor: false
    C:
      small:
        min_detected: 1
        above_floor: false
      large:
        min_detected: 1
        above_floor: false

  intensity_floor_percentile: 10

# -----------------------------------------------------------
# Protein Group Handling
# -----------------------------------------------------------
protein_groups:
  ambiguous_handling: "expand"

# -----------------------------------------------------------
# Plasma Protein Flagging
# -----------------------------------------------------------
plasma_proteins:
  built_in:
    - ALB
    - HBA1
    - HBA2
    - HBB
    - FGA
    - FGB
    - FGG
    - C3
    - C4A
    - C4B
    - C5
    - C6
    - C7
    - C8A
    - C8B
    - C8G
    - C9
    - APOA1
    - APOB
    - APOC1
    - APOC2
    - APOC3
    - APOE
    - TF
    - A2M
    - HP
  patterns:
    - "^IGH"
    - "^IGL"
    - "^IGK"
    - "^KRT"
  external_list: null

# -----------------------------------------------------------
# Contaminant Removal
# -----------------------------------------------------------
contaminants:
  maxquant_prefixes: ["REV__", "CON__"]

# -----------------------------------------------------------
# Peptide Feasibility (Step 5)
# -----------------------------------------------------------
peptide_feasibility:
  missed_cleavages: [0, 1, 2]
  peptide_length_min: 7
  peptide_length_max: 25
  top_n_candidates: 200
  tiers:
    I:
      min_conserved_proteotypic: 2
    II:
      min_human_proteotypic: 1
    III:
      description: "No good proteotypic peptides"

# -----------------------------------------------------------
# Module Validation (Step 6)
# -----------------------------------------------------------
modules:
  dataset: "D11"
  min_completeness: 0.30
  method: "correlation_clustering"
  min_module_size: 20
  enrichment_test: "hypergeometric"
  imputation: "half_min"

# -----------------------------------------------------------
# Sensitivity Analysis (Step 7)
# -----------------------------------------------------------
sensitivity:
  tier_options: ["A", "AB", "ABC"]
  group_handling_options: ["expand", "exclude"]
  weight_sets:
    default:
      mouse_csf: 0.25
      human_csf: 0.30
      ev: 0.05
      brain_plausibility: 0.10
      autophagy_membership: 0.25
    equal:
      mouse_csf: 0.19
      human_csf: 0.19
      ev: 0.19
      brain_plausibility: 0.19
      autophagy_membership: 0.19
    csf_heavy:
      mouse_csf: 0.30
      human_csf: 0.40
      ev: 0.05
      brain_plausibility: 0.05
      autophagy_membership: 0.15
  null_simulation:
    n_iterations: 1000
    proteome_size: 20000
    gene_set_size: 604
    seed: 42

# -----------------------------------------------------------
# Core Panel Selection
# -----------------------------------------------------------
core_panel:
  min_size: 30
  max_size: 80
  require_human_csf_tier: ["A", "B"]
  require_mouse_csf_tier: ["A", "B"]
  require_r1: true
  exclude_plasma: true
  require_ev: false

# -----------------------------------------------------------
# AD Model Cross-Check (Step 8)
# -----------------------------------------------------------
ad_model:
  file: "Table S3 Alterations in proteins identifed in all samples of AppNL-G-F NL-G-F vs Appwt wt.xlsx"
  comparisons:
    - sheet: "AppNL-G-F vs Appwt"
      label: "AppNLGF_vs_WT"
    - sheet: "AppNL-F vs Appwt"
      label: "AppNLF_vs_WT"
    - sheet: "AppNL-F vs AppNL-G-F"
      label: "AppNLF_vs_AppNLGF"
  significance_threshold: 0.05

# -----------------------------------------------------------
# Supplementary Gene Lists (Step 10)
# -----------------------------------------------------------
# Reusable: add new entries here as curated lists grow over time.
supplementary_gene_lists:
  R1b_edison_cosmos:
    file: "novel_candidates_INCLUDE_56_extra_from_Edison_Cosmos_run.xlsx"
    label: "R1b - Edison/Cosmos 2026-02"
    description: "56 novel autophagy genes + 9 HADb core proteins from systematic database search"
    sheets:
      all_candidates: "All Extra Candidates"
      hadb_core: "9 HADb Core - Missing from List"
    gene_symbol_column: "Gene Symbol"
    hadb_category_column: "Category"

# -----------------------------------------------------------
# Output Preferences
# -----------------------------------------------------------
output:
  figure_format: "pdf"
  figure_dpi: 300
  tsv_separator: "\t"
  write_intermediates: true