每一个数字都说明来源和统计边界

Sum of three author-deposited filtered matrices: WT_01 5,321 + WT_02 3,278 + ZC3H20_KO 2,295. The WT trajectory alone contains 8,599 cells.

The catalogue total deliberately includes WT and knockout matrices and is not labelled as the WT-only atlas size. The paper spells the strain EATRO 1125 AnTa1.1 90:13; common AnTat1.1 wording is retained only as an alias.

Fly-derived parasite transcriptomes: 78 midgut + 34 proventriculus + 276 salivary gland. The paper additionally reports 46 in-vitro procyclic single cells used to validate the protocol; those are not included in the catalogue total.

The 388-cell value describes the QC-passing fly-stage atlas rather than every single-cell library reported in the study. J10 is MCRO/ZM/73/J10 and 1738 is MOVS/KE/70/EATRO 1738; DP is a sorting group, not a strain. C5 is predominantly J10 and C6 predominantly 1738, so strain and stage effects may be confounded. The current ENA report exposes 468 runs, while the paper reports 515 sequenced tsetse parasites; these counts are retained separately.

Rows in the author phase tables, excluding headers: 4,366 bloodstream-form + 7,640 procyclic-form cells.

The total was independently reproduced from the two author-deposited cell-level phase tables. The experimental lines are Lister 427-derived. Reads were mapped to T. brucei WT427 2018 and L. major Friedlin references with Cell Ranger 7 so L. major cells and mixed-species multiplets could be removed. BSF and PCF coordinates remain separate because the authors computed them in separate embedding spaces.

Parasite transcriptomes passing quality control: 44 metacyclic forms + 24 at 4 h + 26 at 12 h + 23 at 24 h + 22 at 7 d + 31 bloodstream forms.

The exact kit name is preserved from the Methods and GEO record. The paper's Figure 4 caption uses Smart-seq2/Nextera terminology, so the portal records this nomenclature difference instead of declaring either source wording invalid.

Valid cells retained after author QC across the lifecycle atlas; the paper reports 15,321 genes detected in total.

The experimental strain is Silvio X10/7 A1; Dm28c 2018 is the mapping reference. BioStudies exposes metadata and raw-read links but no processed count matrix. The author atlas serves a processed AnnData view; TrypCell does not present that active scaled matrix as raw counts.

Final retained singlets. The paper first reports 3,192 cells after quality filtering, then removes 41 doublets to retain 3,151 cells for downstream analysis.

AMA/CDT parasites were obtained from infected rat H9c2 myoblasts (ATCC CRL-1446); the SRA repository field “isolation source: Axenic culture” is not treated as the experimental sample origin. The paper and public code use different nUMI/nGene thresholds. TrypCell reports the discrepancy and does not reconstruct a final UMAP because final author cell metadata and embeddings are not deposited.