📄 Methods & data

How the results on this site were produced, and where the underlying samples came from.

RNA-seq pipeline

Tumor RNA-seq was processed with the nf-core/oncoanalyser pipeline (Hartwig tooling) in an RNA-only, tumor-only configuration — there is no matched normal and no DNA sequencing in this analysis, so all DNA/germline modules are intentionally disabled.

FASTQ fastp (trim) STAR (align) Isofox (expression + fusions) CUPPA (tissue of origin)
GTEx comparison: the same two RNA samples were also reprocessed through the GTEx V11 pipeline (STAR → RSEM → RNA-SeQC) so their expression can be compared directly against GTEx normal tissues and PCAWG tumors — see Expression vs GTEx.

Immune repertoire (TCR/BCR)

TRUST4 (Song et al., Nat Methods 2021) was run on the three RNA BAMs to de-novo assemble T-cell and B-cell receptor CDR3 sequences, as an independent cross-check of oncoanalyser's CIDER module (both reconstruct the IG/TR repertoire from the same reads). Candidate reads over the V/D/J/C loci plus unmapped mates are extracted (with --abnormalUnmapFlag, required for these BAMs), assembled, and annotated against IMGT. Clone counts are normalized to library size (per million Isofox fragments) so samples of different depth compare fairly — see Immune repertoire.

Interpretation & limits: bulk RNA-seq samples the repertoire shallowly and is biased toward highly-expressed (plasma) cells, so the low clone overlap between the metastasis and normal lung is largely sampling rather than biology, and the "met-enriched" candidate clones are hypothesis-generating leads — sequence alone cannot establish that an antibody recognizes the tumor.

Sequencing quality control

Per-sample RNA-seq quality metrics from Isofox; both tumor samples pass QC.

SampleQCTotal fragments DuplicatesSpliced genes Read lengthMedian fragment
Left hand (primary) PASS 41,404,403 23.5% 21,354 76 bp 299 bp
Lung metastasis PASS 124,808,404 50.5% 26,729 101 bp 133 bp
Normal lung (baseline) PASS 111,295,463 55.6% 25,638 101 bp 127 bp
On the numbers: the two Treehouse libraries (lung metastasis and normal lung) have higher duplicate rates and shorter fragments than the Tempus primary, consistent with their low-input (100 ng) preparation. All three samples pass and yield tens of thousands of expressed genes. The normal lung sample is an in-patient baseline processed identically to the tumors — a specificity control (for example, the diagnostic EWSR1::KLF15 fusion is present in the tumor but absent in normal lung), not itself a tumor.

Open full MultiQC report ↗

Samples shown on this site

3 RNA-seq samples were reprocessed locally and are shown here — two of the catalogued tumor specimens plus a normal-lung baseline (a within-patient control, not itself a catalogued tumor):

Full sample catalogue

#AccessionSitePhaseTypeDiagnosisFusionNGSVendor
1 SUH-21-014093 Left hand Primary / earliest known specimen FFPE MEC EWSR1::KLF15 WES, RNA-seq, single cell Tempus; NeoGenomics; BostonGene
2 LPS-22-04880 Lung, right lower lobe Pulmonary metastatic / recurrent lesion FFPE MEC Not yet documented N/A N/A
3 CF-01500-2 Lung, left lower lobe Pulmonary metastatic / recurrent lesion Cryopreserved MEC EWSR1::KLF15 WES, RNA-seq BGI Genomics
4 SP-22-081918-C Lung, left lower lobe Pulmonary metastatic / recurrent lesion Originally Frozen MEC EWSR1::KLF15 RNA-seq Covance by Labcorp
5 SP-22-081918-A Lung, right lower lobe Pulmonary metastatic / recurrent lesion FFPE MEC Not yet documented WES, RNA-seq, single cell Tempus; BostonGene
6 SP-22-081918-B Lung, right upper lobe Pulmonary metastatic / recurrent lesion FFPE MEC Not yet documented N/A N/A

Diagnosis abbreviations are as recorded in the clinical sample sheet. Several specimens have whole-exome sequencing (WES) held at external vendors; those data are not part of this analysis.

Contact Us

Please contact us with any questions or comments.