DeepSeek助力快速生成DNBC4tools所需样本对应信息

$sample1 /data/cDNA1_R1.fq.gz;/data/cDNA1_R2.fq.gz /data/oligo1_R1.fq.gz,/data/oligo4_R1.fq.gz;/data/oligo1_R2.fq.gz,/data/oligo4_R2.fq.gz $sample2 /data/cDNA2_R1.fq.gz;/data/cDNA2_R2.fq.gz /data/oligo2_R1.fq.gz;/data/oligo2_R2.fq.gz $sample3 /data/cDNA3_R1.fq.gz;/data/cDNA3_R2.fq.gz /data/oligo3_R1.fq.gz;/data/oligo3_R2.fq.gz

ls *.gz | awk -F_ 'BEGIN {OFS="\t"} {    sample = $1;    lib = $2;    split($3, temp, ".");    direction = temp[1];        all_samples[sample] = 1;        if (lib ~ /^L[0-9]+/) {        lib_num = substr(lib, 2) + 0;  # 提取L后面的数字                # 记录文库顺序        if (!(sample SUBSEP lib_num in lib_seen)) {            libs_order[sample] = libs_order[sample] ? libs_order[sample] " " lib_num : lib_num;            lib_seen[sample, lib_num] = 1;        }                # 存储文件路径        if (direction == "f1") {            f1_files[sample, lib_num] = $0;        } else if (direction == "r2") {            r2_files[sample, lib_num] = $0;        }    } else if (lib == "oligo") {        if (direction == "f1") {            oligo_f1[sample] = $0;        } else if (direction == "r2") {            oligo_r2[sample] = $0;        }    }}END {    for (sample in all_samples) {        # 处理L00数据        l00_f1 = "";        l00_r2 = "";                if (sample in libs_order) {            split(libs_order[sample], lib_nums, " ");            n = length(lib_nums);                        # 冒号排序确保文库顺序            for (i=1; i<=n; i++) {                for (j=1; j<=n-i; j++) {                    if (lib_nums[j+1] < lib_nums[j]) {                        tmp = lib_nums[j];                        lib_nums[j] = lib_nums[j+1];                        lib_nums[j+1] = tmp;                    }                }            }                        # 构建f1和r2列表            for (i=1; i<=n; i++) {                num = lib_nums[i];                if ((sample, num) in f1_files) {                    l00_f1 = l00_f1 ? l00_f1 "," f1_files[sample, num] : f1_files[sample, num];                }                if ((sample, num) in r2_files) {                    l00_r2 = l00_r2 ? l00_r2 "," r2_files[sample, num] : r2_files[sample, num];                }            }        }                # 合并L00字段        l00_combined = l00_f1 ";" l00_r2;                # 处理Oligo数据        oligo_data = "";        if (oligo_f1[sample] && oligo_r2[sample]) {            oligo_data = oligo_f1[sample] ";" oligo_r2[sample];        }                print sample, l00_combined, oligo_data;    }}'  > sample2.tsv

$cat sample2.tsv D10     D10_L001_f1.fq.gz,D10_L002_f1.fq.gz;D10_L001_r2.fq.gz,D10_L002_r2.fq.gz D10_oligo_f1.fq.gz;D10_oligo_r2.fq.gzD15     D15_L001_f1.fq.gz,D15_L002_f1.fq.gz;D15_L001_r2.fq.gz,D15_L002_r2.fq.gz D15_oligo_f1.fq.gz;D15_oligo_r2.fq.gzD5-2    D5-2_L001_f1.fq.gz,D5-2_L002_f1.fq.gz;D5-2_L001_r2.fq.gz,D5-2_L002_r2.fq.gz     D5-2_oligo_f1.fq.gz;D5-2_oligo_r2.fq.gzD2-1    D2-1_L001_f1.fq.gz,D2-1_L002_f1.fq.gz;D2-1_L001_r2.fq.gz,D2-1_L002_r2.fq.gz     D2-1_oligo_f1.fq.gz;D2-1_oligo_r2.fq.gzD8-2    D8-2_L001_f1.fq.gz,D8-2_L002_f1.fq.gz;D8-2_L001_r2.fq.gz,D8-2_L002_r2.fq.gz     D8-2_oligo_f1.fq.gz;D8-2_oligo_r2.fq.gzD5-1    D5-1_L001_f1.fq.gz,D5-1_L002_f1.fq.gz;D5-1_L001_r2.fq.gz,D5-1_L002_r2.fq.gz     D5-1_oligo_f1.fq.gz;D5-1_oligo_r2.fq.gzD2-2    D2-2_L001_f1.fq.gz,D2-2_L002_f1.fq.gz;D2-2_L001_r2.fq.gz,D2-2_L002_r2.fq.gz     D2-2_oligo_f1.fq.gz;D2-2_oligo_r2.fq.gzD12     D12_L001_f1.fq.gz;D12_L001_r2.fq.gz     D12_oligo_f1.fq.gz;D12_oligo_r2.fq.gzD8-1    D8-1_L001_f1.fq.gz,D8-1_L002_f1.fq.gz;D8-1_L001_r2.fq.gz,D8-1_L002_r2.fq.gz     D8-1_oligo_f1.fq.gz;D8-1_oligo_r2.fq.gz

dnbc4tools rna multi --list sample2.tsv --genomeDir ~/reference/human/homo_ensembl_112_dnbc4_index --threads 10