Review Papers Claim by Claim

Visual summary (figures first — read after)

Key visual takeaways

• High base-level accuracy: QV Hap1=64.1, Hap2=61.8 (Merqury) — per-paper benchmarks for reference-quality human assemblies. Merqury QV and completeness
• Assembly spans centromeres for all chromosomes, but centromeres and rDNA remain the major sources of divergence and residual gaps/unassigned contigs. Centromere and rDNA limitations
• The canonical RPE-1 marker t(X;10)(Xq28;10q21.2) is resolved to base-pair precision with a 73.18 Mb chr10q duplication fused to chrX and a 3603 bp microdeletion at the breakpoint, validated by HiFi/ONT read alignments and Hi-C contact maps. t(X;10) breakpoint curation

Critical analysis — strengths

1. Data depth and orthogonality: PacBio HiFi (46×), ONT ultra-long (~125× total; ~30× >100 kb), and Hi-C (~30×) combined with Strand-seq and cytogenetics provide multiple, independent lines of evidence for contiguity, phasing, and structural accuracy — a best-practice multi-platform design. Sequencing design and orthogonal validation
2. State-of-the-art assembly & curation: Use of Verkko for graph-based diploid assembly plus manual Hi-C guided curation and graph/ultra-long read gap patching (ONT patches polished by HiFi/DeepVariant) reduces assembler-specific artefacts. Authors tested Verkko v1.4 vs v2.0 and chose the draft with fewer problematic collapses — good transparency. Assembler selection and curation pipeline
3. Transparent QC and data release: Merqury, CRAQ, Flagger, compleasm reported; raw reads and genome deposited (GenBank, SRA) and code/annotation at GitHub + Zenodo — facilitates reproducibility. QC and data availability

Critical analysis — limitations, blindspots, and caveats

• Residual gaps and rDNA: ~43 Mb of unresolved rDNA and several centromeric low-confidence regions remain unassigned or flagged; users should consult the Flagger low-confidence track before interpreting variants in these loci. This constrains true T2T completeness for some arrays. rDNA and flagged regions
• Haplotype label arbitrariness: Hi-C phasing assigns Hap1/Hap2 arbitrarily per chromosome; authors acknowledge inability to assign maternal/paternal origin without parental trios — important for studies needing parental origin (imprinting, allele-specific epigenetics). Strand-seq supports chromosome-scale phasing but not parental origin. Phasing limitations
• Triplicated/collapsed region complexity: The t(X;10) triplication/duplication created local assembly collapse risk (chr10/chrX regions accounted for large fraction of gaps). Although the authors carefully curated the structural variant, complex CNV regions remain the hardest to assemble confidently; users should treat copy-number and gene-dosage conclusions in the chr10q duplicated region cautiously and consult raw read alignments. Triplex/assembly collapse warning
• Cell-line specificity and generalizability: RPE-1 is a female, non-cancerous, immortalized line with a stable but cell-line–specific rearrangement (t(X;10)) and hTERT plasmid insertion; results are directly applicable to RPE-1 but cannot be assumed to represent primary retinal tissue or other cell lines. The authors correctly call for more cell-line references. Cell-line generalizability note

Practical recommendations for users who work with RPE-1

1. Use RPE1v1.1 (Hap1/Hap2) as the default mapping reference when analyzing sequencing from RPE-1 experiments to reduce mapping bias, especially in centromeres and structural variation analyses; authors provide UCSC hub links and GenBank accessions. Data release and recommendation
2. Before interpreting structural variants or copy-number in chr10q/chrX, validate with read-backed evidence (long reads spanning junctions), and consult the Flagger low-confidence track provided in the GitHub/Zenodo release. Flagger track and read-backed validation
3. For allele-specific studies (imprinting, parental-origin inference) complement with orthogonal phasing (trio, methylation-based phasing or Strand-seq parental-phasing workflows) because Hi-C labels are arbitrary per chromosome. Need for parental phasing

What would falsify the main claims (how to test robustness)

• Independent assembly of RPE-1 from an independent DNA prep + different assembler(s) yielding no t(X;10) translocation or a markedly different breakpoint would falsify the translocation claim; however cytogenetics and Hi-C in the paper already support t(X;10). t(X;10) orthogonal validation
• Re-analyses showing extensive mis-phasing across many chromosomes (contradicting Strand-seq-supported phasing) would challenge the claimed phasing accuracy — repeatable Strand-seq or parental-trio phasing would settle this. Strand-seq phasing validation

Conclusions — balanced

RPE1v1.1 is a high-quality, near–reference-grade diploid assembly for an experimentally crucial human cell line. It will meaningfully reduce reference bias for RPE-1 functional genomics and genome-editing work, and the public data/resources are valuable. Residual centromeric and rDNA uncertainties, arbitrary per-chromosome haplotype labels (no parental information), and assembly complexity at the t(X;10)/chr10q triplicated region are the main caveats; users must interpret copy-number/dosage calls there with read-level confirmation and consult the provided low-confidence tracks.

Interactive actions

Run deeper bioinformatics (one-click AI agent)

Reference (primary)

Volpe et al. 2025 (RPE1v1.1)

Author reviews (one-click)

Short next steps to improve the assembly or downstream use

1. Generate additional ultra-long ONT coverage focused on unassigned rDNA/SAAC regions and re-run graph-based assembly to try to place the remaining 43 Mb of rDNA.
2. Obtain parental DNA (if possible) or apply methylation-based phasing (ONT methylation signal) to generate consistent maternal/paternal labels across chromosomes.
3. Run orthogonal CNV/optical-mapping (Bionano) to resolve complex triplications at chr10q and confirm copy-number breakpoints at high confidence.

Confidence & meta

Primary claims (assembly contiguity, t(X;10) presence, plasmid insertion, QV) are strongly supported by multi-platform data; residual uncertainty resides mainly in repetitive/rDNA arrays and in haplotype parental assignment. For high-stakes genome edits or haplotype-specific functional studies, confirm with orthogonal methods (optical mapping, parental phasing or methylation).