Initial commit: digital-patients pipeline (clean, no large files)

Large reference/model files excluded from repo - to be staged to S3 or baked into Docker images.
This commit is contained in:
2026-03-26 15:15:23 +01:00
commit 9e6a16c19b
45 changed files with 7207 additions and 0 deletions

217
test.nf Normal file
View File

@@ -0,0 +1,217 @@
nextflow.enable.dsl=2
// ================= IMAGES GO HERE ================================================================
params.containerOptions = '--rm' // '--gpus all --rm -v /mnt:/mnt'
params.containerOptions_synthea = "--rm --privileged --gpus all" // -v /mnt:/mnt"
params.containerOptions_borzoi = '--rm --gpus all' // '--gpus all --rm -v /mnt:/mnt'
params.container_borzoi = 'harbor.cluster.omic.ai/omic/digital-patients/borzoi:latest'
params.container_vcf2prot = "harbor.cluster.omic.ai/omic/digital-patients/vcf2prot:latest"
params.container_rna2protexpression = 'harbor.cluster.omic.ai/omic/digital-patients/rna2protexpression:latest'
params.containerOptions_rna2protexpression = '--gpus all --rm' // -v /mnt:/mnt -v /dbs:/dbs'
// params.containerOptions_rna2protexpression = '--gpus all --rm -v /mnt:/mnt -v /dbs:/dbs'
params.container_corto = 'harbor.cluster.omic.ai/omic/digital-patients/corto:latest'
params.container_ecotyper = 'harbor.cluster.omic.ai/omic/digital-patients/ecotyper:latest'
params.container_synthea = 'harbor.cluster.omic.ai/omic/digital-patients/synthea:cudf'
// =================================================================================================
//SYNTHEA
params.n_pat = 2 //10 //number of patients to generate, must be => 2 (one male one female)
params.percent_male = 0.5 // between 0-1, percent of male patients
//params.age = '18-80' //age range of the population, expressed as age-age
//params.state = '"District of Columbia"' //'Hawaii' //'Texas' //'Georgia'
//params.city = '' //'Washington' //'' //'Honolulu' //'Houston' //'Atlanta'
// Keep both versions of disease params
params.disease = 'schizophrenia' //'schizophrenia' //'healthy' //'leukaemia' //'leukaemia' //'Purpura and other haemorrhagic conditions' //['tongue cancer', 'dementia', 'arthritis'] //'tongue cancer' //'dementia'
//params.n_var = 100
params.project_name = 'test'
//CIBERSORT
params.cibersortx_username = "gabriel.richman.2009@anderson.ucla.edu"
params.cibersortx_token = "b5b39e563fb947df4cfd4843d40fdb99"
// ====================== FILEPATHS HERE =========================================================================
params.imputed_store = '/mnt/Avatar/imputed/ukbb/imputed' //'/rosalind/ukbb/imputed'
params.pheno_store ='/mnt/dreamdock-data/digital-patient-data/data/ukbb_phenotypes_filtered.csv' // '/mnt/Avatar/dd/synthea/metadata/ukbb_phenotypes_filtered.csv'
//CORTO
params.regulon = '/mnt/dreamdock-data/digital-patient-data/data/regulon.rda' // '/Workspace/next/registry/pipelines/digital_patient/regulon.rda'
params.healthy_dir = '/mnt/dreamdock-data/digital-patient-data/healthy' // '/Workspace/next/registry/pipelines/digital_patient' //'/mnt/Avatar/digital_patient' // data copy is here
params.synthea_support_dir = '/mnt/Avatar/dd/synthea/supporting-data/'
//BORZOI
params.mane = '/mnt/dreamdock-data/digital-patient-data/data/MANE.GRCh38.v1.3.update.tsv' // '/Workspace/next/registry/pipelines/digital_patient/MANE.GRCh38.v1.3.update.tsv'
//CIBERSORT
params.signature_matrix = '/mnt/dreamdock-data/digital-patient-data/data/LM22_sourceGEP_ensg.txt' // "/Workspace/next/registry/pipelines/digital_patient/LM22_sourceGEP_ensg.txt"
params.outdir = '/mnt/dreamdock-data/digital-patient-data/out' // '/mnt/OmicNAS/dd/digital_patient/new'
// params.outdir = '/data/digital-patients-data' // '/mnt/OmicNAS/dd/digital_patient/new'
params.ecotyper_outdir = "${params.outdir}/ecotyper"
// ====================== FILTERING PARAMETERS ====================================================================
// Enable/disable filtering
params.enable_filtering = true
// Filtering thresholds based on Gabe's requirements
params.top_n_variants = 500
params.top_n_genes = 1000
params.transcriptome_log2fc_threshold = 1.5
params.top_n_proteins = 500
params.top_n_immune_cells = 20
params.top_n_metabolites = 50
params.metabolome_pvalue_threshold = 0.05
// ===============================================================================================================
//SYNTHEA
include {get_disease_stats_no_patients} from './main_synthea.nf'
include {generate_m_variants_cudf} from './main_synthea.nf'
include {generate_f_variants_cudf} from './main_synthea.nf'
include {make_vcfs} from './main_synthea.nf'
include {generate_m_healthy_cudf} from './main_synthea.nf'
include {generate_f_healthy_cudf} from './main_synthea.nf'
//BORZOI
include {FILTER_VCF} from './main_borzoi.nf'
include {PREDICT_EXPRESSION} from './main_borzoi.nf'
include {CREATE_PROTEIN_CLUSTER} from './main_borzoi.nf'
//VCF2PROT
include {VCF2PROT} from './main_vcf2prot.nf'
//RNA2PROTEINEXPRESSION
include {RNA2PROTEXPRESSION} from './main_rna2proteinexpression'
//CORTO
include {CORTO} from './main_corto.nf'
//CIBERSORT
include {CONVERT_TO_TXT} from './main_cibersortx.nf'
include {CIBERSORTx_FRACTIONS} from './main_cibersortx.nf'
include {CIBERSORTx_HIRES} from './main_cibersortx.nf'
include {ADD_TISSUE_NAMES_TO_CIBERSORTX} from './main_cibersortx.nf'
//FILTERING PROCESSES
include {FILTER_VARIANTS} from './main_filter_outputs.nf'
include {FILTER_TRANSCRIPTOME} from './main_filter_outputs.nf'
include {FILTER_PROTEOME} from './main_filter_outputs.nf'
include {FILTER_IMMUNE_CELLS} from './main_filter_outputs.nf'
include {FILTER_METABOLOME} from './main_filter_outputs.nf'
include {FILTER_MUTATED_PROTEINS} from './main_filter_outputs.nf'
include {CREATE_SUMMARY_REPORT} from './main_filter_outputs.nf'
workflow {
pheno_store_ch = file(params.pheno_store)
imputed_store_ch = file(params.imputed_store)
synthea_support_ch = file(params.synthea_support_dir)
regulon_ch = file(params.regulon)
signature_matrix_ch = file(params.signature_matrix)
mane_ch = file(params.mane)
health_dir_ch = file(params.healthy_dir)
//SYNTHEA
switch (params.disease) {
case 'healthy':
//healthy
generate_m_healthy_cudf(health_dir_ch)
generate_f_healthy_cudf(health_dir_ch)
m_healthy = generate_m_healthy_cudf.out
f_healthy = generate_f_healthy_cudf.out
txt_ch = f_healthy.mix(m_healthy).flatten()
break
default:
//disease
get_disease_stats_no_patients(pheno_store_ch, imputed_store_ch)
generate_m_variants_cudf(get_disease_stats_no_patients.out)
generate_f_variants_cudf(get_disease_stats_no_patients.out)
f_var = generate_f_variants_cudf.out
m_var = generate_m_variants_cudf.out
txt_ch = f_var.mix(m_var).flatten()
}
make_vcfs(txt_ch, synthea_support_ch)
// TODO: ADD STEP TO CREATE AVERAGED COHORT (HEALTHY)
// TODO: ADD STEP TO CREATE AVERAGED DISEASED
// TODO: ADD STEP TO CALCULATE STATISTICS LOG2FC and ADJ PVAL
//BORZOI
// Prepare the VCF file(s)
//vcf_ch = Channel.fromPath(make_vcfs.out)
// THIS IS USES TO CREATE PROTEIN CLUSTER ON A SAME BORZOI INPUT AND SPLIT ONES TOO BIG FOR BORZOI (LEN = 524288)
// THIS STEP NEEDS TO BE RUN TO CREATE DATA FOR NEXT STEP ONLY IF NEW REFERENCE/RNAS ARE GOING TO BE USED
//CREATE_PROTEIN_CLUSTER(params.mane)
FILTER_VCF(mane_ch, make_vcfs.out)
PREDICT_EXPRESSION(FILTER_VCF.out, mane_ch)
//VCF2PROT
VCF2PROT(make_vcfs.out, FILTER_VCF.out)
//RNA2PROTEINEXPRESSION
// Updated RNA to protein expression step to run only a single process at a time, avoids CUDA out-of-memory errors.
PREDICT_EXPRESSION.out
.collect()
.flatten()
.set { rna_input }
RNA2PROTEXPRESSION(rna_input)
//CORTO
CORTO(PREDICT_EXPRESSION.out, regulon_ch)
//CIBERSORT
signature_file = Channel.fromPath(signature_matrix_ch, checkIfExists: true)
CONVERT_TO_TXT(PREDICT_EXPRESSION.out)
CIBERSORTx_FRACTIONS(CONVERT_TO_TXT.out, signature_file)
CIBERSORTx_HIRES(CONVERT_TO_TXT.out, CIBERSORTx_FRACTIONS.out, signature_file)
ADD_TISSUE_NAMES_TO_CIBERSORTX(CONVERT_TO_TXT.out, CIBERSORTx_HIRES.out)
// ==================== FILTERING STAGE ====================
if (params.enable_filtering) {
// Filter variants
FILTER_VARIANTS(make_vcfs.out)
// Filter transcriptome
FILTER_TRANSCRIPTOME(PREDICT_EXPRESSION.out)
// Filter proteome
FILTER_PROTEOME(RNA2PROTEXPRESSION.out)
// Filter immune cells
FILTER_IMMUNE_CELLS(ADD_TISSUE_NAMES_TO_CIBERSORTX.out)
// Filter metabolome
FILTER_METABOLOME(CORTO.out)
// Filter mutated proteins
FILTER_MUTATED_PROTEINS(VCF2PROT.out, FILTER_VARIANTS.out.filtered_vcf)
// Create summary report
CREATE_SUMMARY_REPORT(
FILTER_TRANSCRIPTOME.out,
FILTER_PROTEOME.out,
FILTER_IMMUNE_CELLS.out,
FILTER_METABOLOME.out,
FILTER_VARIANTS.out.filtered_vcf
)
}
}