35 of 54 contrast × method slots have a scored model (held out: cells (no donor column)); best held-out AUC 1.000 (T vs B cells · combined (hash + new screen 1, all k-mer)); the combined hash + new screen 1 model beats the hash model in 7 of 9 contrasts

35 models over 9 contrasts, scored on held-out cells (no donor column)

Dataset: bmcite (Seurat bone-marrow CITE-seq, GEO GSE128639) · labels from the Seurat reference annotation · 2 SRA runs · 30,672 cells in the metadata · 2 donors. Counted over: every contrast × method with an output under the S3 prefixes sent for this page. Each contrast keeps only the cells of its classes. There is no donor column, so P2G holds out random cells: the score is on new cells from the same people.

Each row is one model. hash is the P2G stage 3 model; new screen 1 ranks targets by presence ratio between the classes; combined fits one lasso on the union of both candidate sets; combined dict averages an all-k-mer model and a SPLASH model. Binary contrasts report held-out AUC; multiclass contrasts report macro AUC (the mean one-vs-rest AUC over all classes); regressions report Spearman ρ (rank agreement of predicted and measured values) and R² (share of the variance explained compared with predicting the mean).

Stuart et al. 2019 reports no held-out AUC for these contrasts

Reference cell-type labels (celltype_l1 / l2, and all 27 WNN types), plus 11 clusters made from surface protein only; there is no donor split

Counted over: the cells of each contrast, joined to the exact train and test cell IDs in each run's run_split.rds. Without a donor column the held-out cells are random cells from the same donors, so the score is optimistic. The two 9 Oct contrasts are multiclass: one model predicts all classes, and each selected sequence is tagged with the class it points to most. The 27 WNN types use RNA and protein together, so they are not protein clusters.

Stage 1, all anchors: 2 to 15,299 of about 53,000 anchors pass |r| > 0.2

Counted over: every anchor in each hash run's stage1_pearson_all.tsv.gz, before any selection. Passes per contrast: T vs B cells 261 · CD4 vs CD8 T cells 2 · 11 protein clusters (multiclass) 8,717 · 27 WNN cell types (multiclass) 15,299 · CD14 vs CD16 monocytes 572 · HSC vs LMPP/GMP progenitors 427 · CD8 naive vs effector/memory 117 · MAIT vs other CD8 T cells 56 · CD4 naive vs memory 18. r is computed on read count × f, so its sign follows f; the page shows |r|. An anchor's gene is the gene of its top target in stage2_cv_pass.tsv.gz.

New screen 1 keeps targets present in enough cells, ranks them by presence ratio, and picks the cut K by training cross-validation

Counted over: every anchor-target pair in each contrast's new screen 1 run, read from its run.log. Exclusive targets are present in one class only; graded targets are present in both and ranked.

Selected sequences: 18,781 of 28,588 selected rows (summed over the 35 models) match the genome exactly

Counted over: the sequences each model selected (non-zero weight): 28,588 across the 35 models, out of 45,023 rows in their tables. 5,889 of them have no target recorded.

Selected genes vs expression: how much of each model's weight sits on highly expressed genes

Counted over: the genes the selected sequences are drawn on, per model, weighted by summed |weight|. Expression: GSM3681518_MNC_RNA_counts.tsv.gz (GEO GSE128639), normalised to counts per 10,000 per cell.

Literature: what PubMed says about each heavy gene with T-cell vs B-cell identity and CD4 vs CD8 lineage

Counted over: the 15 heaviest named genes of each model. PubMed title/abstract searches: gene + T-cell vs B-cell identity, gene + CD4 vs CD8 lineage, and gene + both (a paper that names the gene and both topics).

144 alternative-splicing, 1,358 mismatch (1 or more) and 269 BLAST-only features (5 of them imperfect) among the selected sequences of the 35 models

Counted over: the selected sequences (non-zero weight) of every model, after the 1% rule: a sequence present in fewer than 1% of training cells is left out (731 selected sequences, almost all from hash models; new screen 1 already requires at least 5%). (a) crosses a STAR junction that is not in the reference transcript; compatible transcripts have that exact intron, incompatible ones span it but splice differently or keep it. (a) also counts a soft-clipped sequence that is full length only on alternative isoforms. (b) edit distance of 1 or more from the aligner that placed it, on a full-length placement (split into 1 and 2 or more); soft-clipped and unplaced sequences have no comparable edit distance and are left out of (b). (a) or (b) is counted once per sequence, with the overlap shown, over the selected sequences and over all stage 3 rows. This follows section 3h of the GSE123813 instructions (Julia, 2026-10-08). (c) placed by neither aligner, but hit by local BLAST (GENCODE, genome, IPD-IMGT/HLA) or NCBI core_nt; perfect = 100% coverage and 100% identity. A sequence can be in more than one group. The same features for every study are collected on the running feature ledger.

Each selected sequence drawn on its gene, marked by how it matches the genome and how other isoforms read it

Drawn: the selected sequences with a genome position. Listed in the table under the track: every row on that gene, selected or not.

Choose a model, then a gene. Genes are listed by the total weight of their selected sequences. Click a mark or a table row for its card.

1,716 selected sequences (summed over the 35 models) did not align confidently and were BLASTed locally

Counted over: selected sequences of the chosen model (section 4's model menu). The pool is every sequence without a confident full-length alignment with at most 1 mismatch; sequences with no target recorded are counted but cannot be searched.

Every sequence with 2+ mismatches or soft-clipping, BLASTed against mRNAs, the whole genome and HLA alleles

Counted over: every row with 2 or more mismatches (full length) or any soft-clipping, selected or not. Coverage = aligned bases ÷ sequence length; imperfect values are amber.

5c · Selected sequences never BLASTed, and the ones at a known SNP or with a base change

Counted over: the selected sequences of the model chosen in section 4.

Selected sequences at a known dbSNP site, a ClinVar or COSMIC record, or with a base change called by P2G

Counted over: the selected sequences of the model chosen in section 4 (heaviest 60). Known variants in HLA and Fc-gamma genes, for every model, are in section 6c.

How much of each model's weight sits on textbook T, B, CD4 and CD8 lineage genes?

Counted over: the selected sequences of every model, by summed |weight|. Groups: textbook T-cell lineage genes, B-cell lineage genes, and CD4 / CD8 lineage genes. The lookup counts every row, selected or not. HLA and Fc-gamma genes, which differ between people, are counted per model in section 6c.

Immune genes that vary between people: how much of each model’s score is in HLA and Fc-gamma genes?

Why this matters: a model can learn which person a cell came from instead of which condition it is in, and HLA is the most variable region of the human genome. This study has no donor column, so test cells can come from the same people as training cells, and this page cannot check the risk.

6 of 35 models are flagged: HLA and Fc-gamma genes make up 5% or more of their score.

8 models select no sequence in these genes: CD4 vs CD8 T cells · hash all k-mer; CD4 vs CD8 T cells · new screen 1 all k-mer; CD4 vs CD8 T cells · combined (hash + new screen 1, all k-mer); 27 WNN cell types (multiclass) · hash SPLASH; 27 WNN cell types (multiclass) · new screen 1 SPLASH; CD14 vs CD16 monocytes · new screen 1 SPLASH; HSC vs LMPP/GMP progenitors · new screen 1 SPLASH; CD8 naive vs effector/memory · new screen 1 SPLASH.

The selected sequences in these genes, model by model

Click a model to open its sequences. 27 models have at least one.

Checks run on this build (they inform, they never block)

Counted over: every model table read for this page.

Annotation coverage per model (rows with a value)

Every file this page was built from, with its S3 path

Read: the files below, the GEO expression matrix and guide barcodes, GENCODE v46, GRCh38.p13, IPD-IMGT/HLA 3.65.0, NCBI gene_info and MSigDB hallmark sets.

Every rule this page followed

Source

Built to match the GSE123813 sequence-review page (artifact 6ChpFqMvKXfBMEbrJT1LsV) and the page instructions in rosabiocom/rosa-skills (population-comparison-site, *_PAGE_INSTRUCTIONS.md). The original build scripts are not in the skill, so this page was rebuilt: Julia's page code (renderer, gene_track_annot.js, styles) is reused, and every data file is regenerated by the shared build in website-artifact/sites/study-build (build_rows.py, run_blast.sh, blast_classify.py, expr_stats.py, lit_fetch.py, build_page.py, assemble.py).

Models

One page per study, every contrast on it. Each contrast has up to six models, chosen in every dropdown: hash all k-mer and hash SPLASH (P2G stage 3, stage3_annotated_target.tsv.gz of each dictionary), new screen 1 all k-mer and new screen 1 SPLASH (presence-ratio final-selected tables), combined (Combined_selected_seq: hash + new screen 1 candidates in one lasso) and combined dict (all-k-mer + SPLASH dictionaries, equal-weight). A slot without an output is listed in section 1 with the reason.

Scores

Binary contrasts: held-out AUC (hash: roc_auc of test_evaluation.tsv at feature_level = anchor_target; the anchor-level AUC of summary_row.tsv is shown for reference). Regressions: Spearman ρ between predicted and measured values on held-out cells, and R², the share of the variance explained compared with predicting the mean. Multiclass contrasts (one model over all clusters or cell types): macro AUC, the mean of the one-vs-rest AUCs over all classes (hash: macro_auc of test_evaluation.tsv; new screen 1: the __macro__ row of New_presratio_per_class_auc; combined and combined dict: macro_auc of their metrics tables).

Weight and direction

hash: lasso_coef × f; the other three: the model coefficient. Positive pushes toward the first class of the contrast, negative toward the second. For a regression, positive raises the predicted value and negative lowers it. For a multiclass contrast each sequence is tagged with the class it points to most (its largest positive per-class coefficient; for hash, the anchor-level coefficient × f) and its weight is shown without a sign.

Rows

One aligner per sequence (minimap2 when it placed the sequence, otherwise bowtie2); CIGAR, edit distance and cs all from that aligner. A dotted or empty cell is "none". Gene = GENCODE v46 gene overlapping the drawn position, protein-coding before lncRNA before pseudogene, named before ENSG-only. Rows with no target sequence are "no target recorded" and never placed or searched. SPLASH features of the combined-dictionary model take their sequence from the SPLASH new screen 1 table when it has one.

Isoforms

GENCODE v46 comprehensive. The reference row is MANE Select, else Ensembl canonical. An isoform row is kept only when it reads a selected sequence differently (stopping short is not a different reading); identical readings are collapsed; at most 8 rows; every transcript is accounted for in the caption.

Junctions

STAR star_junction matched exactly to the introns of every GENCODE v46 transcript: reference, annotated alternative, or not annotated (new pairing, one new site, no annotated site).

BLAST

Local first: GENCODE v46 transcripts (blastn-short, dust off, E ≤ 1, ≥ 90% identity, up to 500 targets), GRCh38.p13 top level with alternate haplotypes and patches (word size 11, E ≤ 10, ≥ 95% identity) and IPD-IMGT/HLA 3.65.0. Pool = no confident full-length alignment with ≤ 1 mismatch. Rescue bar: ≥ 30 bases at ≥ 97% identity; transcripts first, then the genome (E ≤ 1e-4). A tie across genes is ambiguous; a genome locus is a position, never a gene. Every sequence no local search explains, and every 5b "no closer match" or "clipped bases unexplained" sequence, was then searched remotely in NCBI core_nt (megablast, word size 16, E ≤ 1e-3, up to 25 targets); the card shows the best hit, its organism and the tie count. Other pool sequences were not sent remotely; a sequence still unsearched would be marked "would need a remote search".

Expression

GEO GSM3681518_MNC_RNA_counts.tsv.gz (GEO GSE128639), CP10K per cell. Each model's two groups are its own train + test cells from run_split.rds (a regression is split at the median of the response). Matrix columns are matched to P2G cells by barcode. Ranks are among genes expressed in the two groups together; a GENCODE name missing from the matrix is matched through NCBI synonyms.

Literature

PubMed title/abstract counts for gene + T-cell vs B-cell identity, gene + CD4 vs CD8 lineage, and both, for the 15 heaviest named genes of each model (symbol, full name, NCBI synonyms of 4+ characters with a digit or hyphen). Up to three papers per gene with the abstract sentence that names the gene, verbatim. Each linked paper has a one-line conclusion written from its title and abstract only, with a relevance tag (central, finding, gene list only, unrelated match, no abstract).

Gene census

Groups: textbook T-cell lineage genes, B-cell lineage genes, and CD4 / CD8 lineage genes. The lookup counts every row, selected or not.

Literature: splicing, mismatches and BLAST per gene

Requested 2026-10-08. For each of the 15 heaviest genes in the literature table (3f), two columns summarise that gene's selected sequences (non-zero weight) in the chosen model: alternative splicing (STAR junctions classed against GENCODE v46, plus sequences exonic only in non-reference isoforms and the mRNA-search junction verdict) and mismatches · BLAST (edit-distance strata 0 / 1 / 2+, soft-clipped and unplaced counts, the local BLAST verdicts, and NCBI core_nt hits). A sequence belongs to the gene when it lies in the gene's track, or its stage-3 gene_name, junction annotation or BLAST transcript hit names the gene.

Feature report (section 3g) and the running ledger

Requested 2026-10-08. Every rebuild runs feature_report.py on the selected sequences (non-zero weight) of every model and keeps the ones in three groups: (a) alternative splicing, a STAR junction not in the reference transcript, with the GENCODE v46 transcripts compatible with it (that exact intron) and incompatible (span it, splice differently or keep it); (b) 1 or more mismatches from the aligner that placed the sequence; (c) no aligner placement but a BLAST hit (local, then NCBI core_nt), perfect or imperfect. The 1% rule is applied: a sequence present in fewer than 1% of training cells is left out (hash rows: Train_prev weighted by class size; new screen 1 already requires 5%; combined rows take the share of the same sequence in the hash or new screen 1 table). build_ledger.py then collates every study into the running feature ledger with study, phenotype, label basis, model, gene, direction, target rank, coefficient, weight and share of cells.

Not done

Cell-type marker status and LitVar2 variant papers were not run for this build. Never downloaded raw reads.

Published study

Stuart et al. 2019 (PMID 31178118) reports no held-out AUC for T vs B or CD4 vs CD8.