data/
├── c315
│   ├── allprot.fasta
│   └── * (*: patient id)
│       ├── proteome.fasta
│       └── peptides_HLA-I.csv
├── c442
│   └── (c315와 동일한 구조로 생성됨)
├── patient_id.txt
└── common_mhc.txt

#!/bin/bash

# 입력:
# 1) 클러스터명 (예: c315)
# 2) 병렬 프로세스 수 (NUM_PROC)
# 출력:
# 환자별 binding_affinities_HLA-I.csv

CLUSTER=$1
NUM_PROC=$2

netMHCpan="../netMHCpan-4.1/netMHCpan"
OUT_DIR="data/${CLUSTER}"
PATIENT_TXT="data/patient_id.txt"
HLA_I_ALLELES_FILE="data/common_mhc.txt"

# 스크립트가 있는 디렉터리로 이동
cd "$(dirname "$0")"

# 환자별로 netMHCpan 예측 수행
while read -r PATIENT_ID; do
    PATIENT_DIR="$OUT_DIR/$PATIENT_ID"
    RAW_DIR="$PATIENT_DIR/raw_predictions"
    
    mkdir -p "$RAW_DIR"

    PEPTIDES_TABLE="$PATIENT_DIR/peptides_HLA-I.csv"
    
    # peptides.txt 생성 (첫 컬럼만 추출, 헤더 제거)
    tail -n +2 "$PEPTIDES_TABLE" | cut -d"," -f1 > "$PATIENT_DIR/peptides.txt"

    # 병렬로 netMHCpan 실행
    parallel -j "$NUM_PROC" \
        "$netMHCpan -BA -p $PATIENT_DIR/peptides.txt -a {} > $RAW_DIR/{}.txt" \
        :::: "$HLA_I_ALLELES_FILE"

    # 결과 병합 및 정리
    python3 sc2.py "$RAW_DIR" > "$PATIENT_DIR/binding_affinities_HLA-I.csv"
    rm -rf "$RAW_DIR" "$PATIENT_DIR/peptides.txt"

done < "$PATIENT_TXT"

import sys
import os

netMHCpan_outdir = sys.argv[1]
print("Allele,Peptide,Affinity")
for fname in sorted(os.listdir(netMHCpan_outdir)):
    netMHCpan_file = open(netMHCpan_outdir + "/" + fname, "r")
    table_started = False
    for line in netMHCpan_file:
        if "Pos" in line:
            table_started = True
            netMHCpan_file.readline()
            continue
        
        if table_started and line[0] == "-":
            break
        
        if not table_started:
            continue

        line = line.split()
        allele = line[1]
        peptide = line[2]
        affinity = int(float(line[15]))
        print("{},{},{}".format(allele, peptide, affinity))
    
    netMHCpan_file.close()

data/
├── c315
│   ├── allprot.fasta 
│   └── * (*: patient id)
│       ├── proteome.fasta
│       ├── peptides_HLA-I.csv
│       └── binding_affinities_HLA-I.csv
├── c442
│   └── (c315와 동일한 구조로 생성됨)
├── patient_id.txt
└── common_mhc.txt