kfukuta/wf-transcriptomes-v202
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Skip cohort processes when n=1 [CW-7208]

Browse Source
This commit is contained in:
Kiah McIntosh 2026-05-21 16:19:18 +00:00
parent d520a7635b
commit 759cb9d868
5 changed files with 216 additions and 40 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
.gitlab-ci.yml
View File

@ -134,15 +134,14 @@ docker-run:
jq -e '.contrasts["condition_treated_vs_control"].dge_status | IN("SUCCESS", "FAILED")' ${CI_PROJECT_NAME}/de_analysis/de_qc_stats.json >/dev/null &&
jq -e '.contrasts["condition_treated_vs_control"].dtu_status | IN("SUCCESS", "FAILED")' ${CI_PROJECT_NAME}/de_analysis/de_qc_stats.json >/dev/null
# Smoke: quick discover-mode sanity check for core cohort and per-sample outputs.
# Smoke: quick discover-mode sanity check for core cohort outputs.
- if: $MATRIX_NAME == "smoke_discover"
variables:
NF_BEFORE_SCRIPT: ":"
NF_WORKFLOW_OPTS: "--fastq test_data/smoke/reads.fastq --sample sampleA --ref_genome test_data/smoke/reference.fa --ref_annotation test_data/smoke/annotation.gtf"
AFTER_NEXTFLOW_CMD: >
test -f ${CI_PROJECT_NAME}/cohort/transcripts.gtf &&
test -f ${CI_PROJECT_NAME}/cohort/cohort.transcriptome.fa &&
test -f ${CI_PROJECT_NAME}/samples/sampleA/transcripts.gtf
test -f ${CI_PROJECT_NAME}/samples/sampleA/transcripts.gtf &&
test -f ${CI_PROJECT_NAME}/samples/sampleA/sampleA.transcriptome.fa
# Smoke: fixed-annotation path sanity check for quantification outputs.
- if: $MATRIX_NAME == "smoke_fixed"
@ -150,8 +149,8 @@ docker-run:
NF_BEFORE_SCRIPT: ":"
NF_WORKFLOW_OPTS: "--fastq test_data/smoke/reads.fastq --sample sampleA --ref_genome test_data/smoke/reference.fa --ref_annotation test_data/smoke/annotation.gtf --transcriptome_mode fixed_annotation"
AFTER_NEXTFLOW_CMD: >
test -f ${CI_PROJECT_NAME}/cohort/transcripts.gtf &&
test -f ${CI_PROJECT_NAME}/cohort/transcript_counts.tsv
test -f ${CI_PROJECT_NAME}/samples/sampleA/transcripts.gtf &&
test -f ${CI_PROJECT_NAME}/samples/sampleA/transcript_counts.tsv
# Smoke: direct-RNA alignment profile and downstream SQANTI output presence.
- if: $MATRIX_NAME == "smoke_direct_rna"
@ -160,7 +159,7 @@ docker-run:
NF_WORKFLOW_OPTS: "--fastq test_data/smoke/reads.fastq --sample sampleA --ref_genome test_data/smoke/reference.fa --ref_annotation test_data/smoke/annotation.gtf --direct_rna"
AFTER_NEXTFLOW_CMD: >
test -f ${CI_PROJECT_NAME}/cohort/alignments/sampleA/reads.bam &&
test -f ${CI_PROJECT_NAME}/cohort/sqanti_cohort/classification_summary.tsv
test -f ${CI_PROJECT_NAME}/samples/sampleA/sampleA_sqanti/classification_summary.tsv
# Smoke: end-to-end DE/DTU wiring and expected contrast output files.
- if: $MATRIX_NAME == "smoke_de"

55
bin/workflow_glue/report.py
View File

@ -68,9 +68,9 @@ def _format_ratio_value(value):
return f"{numeric:.2f}x"
def _cohort_summary(cohort_dir):
"""Return cohort-level metrics and transcript class counts DataFrames."""
tx_meta = _read_table(Path(cohort_dir) / "transcript_metadata.tsv")
def _transcriptome_summary(transcriptome_dir):
"""Return transcriptome metrics and transcript class counts DataFrames."""
tx_meta = _read_table(Path(transcriptome_dir) / "transcript_metadata.tsv")
if tx_meta is None:
return None, None
@ -95,7 +95,10 @@ def _cohort_summary(cohort_dir):
def _sample_summaries(samples_dir):
"""Return a dict of per-sample metrics DataFrames keyed by sample name."""
summaries = {}
for sample_dir in sorted(Path(samples_dir).iterdir()):
samples_path = Path(samples_dir)
if not samples_path.exists() or not samples_path.is_dir():
return summaries
for sample_dir in sorted(samples_path.iterdir()):
if not sample_dir.is_dir():
continue
tx_meta = _read_table(sample_dir / "transcript_metadata.tsv")
@ -143,9 +146,9 @@ def _contrast_results(de_dir, filename, n=None):
return tables
def _load_bambu_qc(cohort_dir):
def _load_bambu_qc(bambu_dir):
"""Load bambu QC statistics JSON."""
qc_file = Path(cohort_dir) / "bambu_qc_stats.json"
qc_file = Path(bambu_dir) / "bambu_qc_stats.json"
if qc_file.exists():
with open(qc_file) as f:
return json.load(f)
@ -345,8 +348,6 @@ def main(args):
use_index=False,
)
# Load bambu QC statistics
bambu_qc = _load_bambu_qc(args.cohort_dir)
annotation_reference_summary = _load_annotation_reference_summary(args.cohort_dir)
if annotation_reference_summary:
@ -437,6 +438,16 @@ def main(args):
h4("Unstranded Annotation Examples")
pre("\n".join(examples))
# Setup for using cohort or single sample bambu results
is_single_sample = len(metadata) == 1
primary_label = metadata[0]["alias"] if is_single_sample else "Cohort"
# Load bambu QC statistics
bambu_dir = (
Path(args.samples_dir) / metadata[0]["alias"] if is_single_sample
else args.cohort_dir
)
bambu_qc = _load_bambu_qc(bambu_dir)
# Add Bambu QC section with warnings
if bambu_qc:
with report.add_section("Bambu Quality Control", "Bambu QC"):
@ -556,29 +567,39 @@ def main(args):
use_index=False,
)
with report.add_section("Cohort transcriptome", "Cohort"):
cohort_metrics, cohort_classes = _cohort_summary(args.cohort_dir)
if cohort_metrics is not None:
with report.add_section(
f"{primary_label} transcriptome",
f"{primary_label} transcriptome"
):
transcriptome_metrics, transcriptome_classes = _transcriptome_summary(bambu_dir)
if transcriptome_metrics is not None:
DataTable.from_pandas(
cohort_metrics,
transcriptome_metrics,
paging=False,
searchable=False,
use_index=False,
)
if cohort_classes is not None:
if transcriptome_classes is not None:
DataTable.from_pandas(
cohort_classes,
transcriptome_classes,
paging=False,
searchable=False,
use_index=False,
)
tx_counts = _read_table(Path(args.cohort_dir) / "transcript_counts.tsv")
tx_counts = _read_table(Path(bambu_dir) / "transcript_counts.tsv")
if tx_counts is not None and not tx_counts.empty:
p("Top transcript rows from the cohort abundance table.")
p(
"Top transcript rows from the "
f"{'sample' if is_single_sample else 'cohort'} abundance table."
)
DataTable.from_pandas(tx_counts.head(20), use_index=False)
with report.add_section("Per-sample transcriptomes", "Per sample"):
if not is_single_sample:
with report.add_section(
"Per-sample transcriptomes",
"Per-sample transcriptomes"
):
tabs = Tabs()
for sample, summary_df in _sample_summaries(args.samples_dir).items():
with tabs.add_tab(sample):

165
bin/workflow_glue/tests/common/test_report.py
View File

@ -277,8 +277,12 @@ def test_report_main_handles_degenerate_bambu_qc_and_read_summary(
}
),
)
samples = tmp_path / "samples"
samples.mkdir()
sample_a = samples / "sampleA"
sample_a.mkdir()
_write(
cohort / "bambu_qc_stats.json",
sample_a / "bambu_qc_stats.json",
json.dumps(
{
"samples": 1,
@ -298,10 +302,6 @@ def test_report_main_handles_degenerate_bambu_qc_and_read_summary(
),
)
samples = tmp_path / "samples"
samples.mkdir()
(samples / "OPTIONAL_FILE").touch()
sqanti = tmp_path / "sqanti"
sqanti.mkdir()
(sqanti / "OPTIONAL_FILE").touch()
@ -346,6 +346,161 @@ def test_report_main_handles_degenerate_bambu_qc_and_read_summary(
)
def test_report_main_uses_cohort_bambu_qc_for_multi_sample_inputs(
monkeypatch,
tmp_path,
):
"""Multi-sample runs should render bambu QC from cohort-level outputs."""
tables = []
monkeypatch.setattr(report.labs, "LabsReport", _FakeReport)
monkeypatch.setattr(report, "Tabs", _FakeTabs)
monkeypatch.setattr(report, "p", lambda *args, **kwargs: None)
monkeypatch.setattr(report, "pre", lambda *args, **kwargs: None)
monkeypatch.setattr(report, "_create_warning_banner", lambda *args, **kwargs: None)
monkeypatch.setattr(report.fastcat, "SeqSummary", lambda *args, **kwargs: None)
monkeypatch.setattr(
report.DataTable,
"from_pandas",
staticmethod(lambda table, *args, **kwargs: tables.append(table.copy())),
)
metadata = _write(
tmp_path / "metadata.json",
json.dumps(
[
{"alias": "sampleA", "has_stats": False},
{"alias": "sampleB", "has_stats": False},
]
),
)
params = _write(tmp_path / "params.json", "{}")
versions = tmp_path / "versions"
versions.mkdir()
_write(versions / "versions.txt", "tool,1.0\n")
cohort = tmp_path / "cohort"
cohort.mkdir()
_write(
cohort / "bambu_qc_stats.json",
json.dumps(
{
"samples": 2,
"library_sizes": {"sampleA": 1200, "sampleB": 900},
"min_library_size": 900,
"max_library_size": 1200,
"median_library_size": 1050,
"library_size_ratio": 1.3333,
"total_transcripts_before_filter": 100,
"total_transcripts_after_filter": 80,
"transcripts_filtered": 20,
"median_transcripts_detected": 70,
"total_genes_after_filter": 60,
"transcriptome_mode": "discover",
"ndr_used": 0.1,
}
),
)
samples = tmp_path / "samples"
samples.mkdir()
sample_a = samples / "sampleA"
sample_a.mkdir()
_write(
sample_a / "bambu_qc_stats.json",
json.dumps({"samples": 1, "library_sizes": {"sampleA": 5}}),
)
_write(
sample_a / "transcript_metadata.tsv",
"TXNAME\tGENEID\n"
"tx1\tgene1\n"
"tx2\tgene1\n"
"tx3\tgene2\n",
)
sample_b = samples / "sampleB"
sample_b.mkdir()
_write(
sample_b / "bambu_qc_stats.json",
json.dumps({"samples": 1, "library_sizes": {"sampleB": 7}}),
)
_write(
sample_b / "transcript_metadata.tsv",
"TXNAME\tGENEID\n"
"txA\tgeneA\n"
"txB\tgeneB\n",
)
sqanti = tmp_path / "sqanti"
sqanti.mkdir()
(sqanti / "OPTIONAL_FILE").touch()
alignment_stats = tmp_path / "alignment_stats"
alignment_stats.mkdir()
(alignment_stats / "OPTIONAL_FILE").touch()
out_report = tmp_path / "wf-transcriptomes-report.html"
args = report.argparser().parse_args(
[
str(out_report),
"--metadata",
str(metadata),
"--alignment_stats_dir",
str(alignment_stats),
"--cohort_dir",
str(cohort),
"--samples_dir",
str(samples),
"--sqanti_dir",
str(sqanti),
"--versions",
str(versions),
"--params",
str(params),
]
)
report.main(args)
assert out_report.exists()
assert any(
"Samples analyzed" in table.to_string() and "2" in table.to_string()
for table in tables
)
assert any(
"Library size ratio (max/min)" in table.to_string()
and "1.33x" in table.to_string()
for table in tables
)
assert any(
"Sample" in table.columns
and "Library Size" in table.columns
and {"sampleA", "sampleB"}.issubset(set(table["Sample"].tolist()))
and "1,200" in table.to_string()
and "900" in table.to_string()
for table in tables
)
per_sample_metric_tables = [
table
for table in tables
if list(table.columns) == ["Metric", "Value"]
and set(table["Metric"].tolist()) == {"Transcripts", "Genes"}
]
assert any(
set(zip(table["Metric"], table["Value"])) == {
("Transcripts", 3),
("Genes", 2),
}
for table in per_sample_metric_tables
)
assert any(
set(zip(table["Metric"], table["Value"])) == {
("Transcripts", 2),
("Genes", 2),
}
for table in per_sample_metric_tables
)
def test_report_main_renders_statistical_methods_and_warnings(
monkeypatch,
tmp_path,

4
main.nf
View File

@ -39,7 +39,7 @@ process makeReport {
tuple val(metadata), path(stats, stageAs: "stats_*")
path "versions/*"
path "params.json"
path cohort_dir, stageAs: "cohort"
path cohort_dir, stageAs: "cohort/*"
path sample_dirs, stageAs: "samples/*"
path sqanti_dirs, stageAs: "sqanti/*"
path de_files
@ -198,7 +198,7 @@ workflow pipeline {
report_input,
software_versions,
workflow_params,
transcriptome.joint_dir,
transcriptome.joint_dir.ifEmpty(OPTIONAL_FILE),
sample_dirs_for_report,
sqanti_dirs_for_report,
de_dir,

1
subworkflows/transcriptome.nf
View File

@ -222,6 +222,7 @@ workflow transcriptome_analysis {
joint_discover = runJointBambuDiscover(
alignments
.toSortedList { a, b -> a[0].alias <=> b[0].alias }
.filter { rows -> rows.size() > 1 }
.map { rows ->
// transform [meta, bam, bai] rows to
// [meta, [alias1...aliasN], [bam1...bamN], [bai1...baiN], sample_sheet]