{"id":"719fad9c-1e1d-47e5-afd3-49fe379f7786","arxiv_id":"2506.00692","paper_version":1,"verdict":"UNVERDICTED","confidence":"HIGH","novelty_score":3.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A literature review that organizes recent SMC-based methods into three classes: joint parameter inference, lineage partitioning, and ARG inference, emphasizing applications in ecology, conservation, and phylogenetics.","lead":"This paper is a review of recent extensions to the sequentially Markovian coalescent (SMC), the statistical framework behind PSMC and MSMC, covering inference of gene flow, life-history traits, structural variants, and full ancestral recombination graphs. It is aimed at ecologists, conservation biologists, and phylogeneticists who want to know which SMC-based tools can be applied beyond effective population size estimation.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified; only editorial gap: 'dispersal rates' in the Abstract lacks a supporting section.","rationale":"The manuscript is a review, so the accept/reject frame is not appropriate; the reader's UNVERDICTED classification is correct. The weakest assumption identified by the reader (maturity of preprint methods) is reasonable but not load-bearing in my assessment: the review is a survey, not a recommendation engine, and it explicitly acknowledges simulation-only validation and user-side risks in Section 6. The strongest claim is descriptive and would fail only if cited methods were systematically misrepresented; I found no such failure. The only unfulfilled promise is the Abstract's 'dispersal rates,' which is a scope-error in a list, not a threat to the review's central synthesis. I therefore leave the reader's verdict unchanged, with a suggested editorial correction.","tokens_in":27749,"tokens_out":12203,"duration_ms":122194,"concrete_test":"Run a keyword scan for 'dispersal' and map each capability named in the Abstract (IM, multi-species coalescent, selfing/dormancy, dispersal, computational advances) to a section. If 'dispersal rates' has no supporting method description, revise the Abstract to say 'migration/gene-flow rates' or remove the item. This editorial check would not change the verdict.","verdict_should_be":"UNCHANGED","load_bearing_attack":"I read the review as a synthesis whose central claim is that recent SMC extensions can target parameters beyond Ne. For that claim to hold, the cited methods must exist, be SMC-based, and estimate the quantities advertised. The paper gives section-level support for isolation-with-migration (§3.2), the multi-species coalescent (§3.3), selfing and dormancy (§2.2), structural variation (§3.1), and ARG reconstruction (§4). Section 6 explicitly concedes that validation is often simulation-based, that natural complexity may not be captured, and that users must validate implementation choices. I did not find an internally inconsistent or unsupported step that would overturn the central synthesis. The one concrete discrepancy is editorial: the Abstract and Introduction list 'inference of dispersal rates' as an advance, but no section describes a dispersal-rate estimator; the closest content is migration and gene-flow inference in §3.2. This mismatch affects one list item, not the load-bearing structure of the review.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper is a review of recent extensions to the sequentially Markovian coalescent (SMC) framework, arguing that SMC-based methods are no longer limited to inferring historical effective population sizes. It provides an SMC/SMC′ primer with a formal algorithm, then surveys three classes of extensions: joint parameter inference (recombination rate, mutation rate, selfing, dormancy, selection), lineage partitioning (inversions, population structure, isolation-with-migration, multi-species coalescent, speciation), and ARG-based inference (ARGweaver, SINGER, Relate, tsinfer, ARG-Needle, Threads). It closes with empirical considerations, data requirements, and future directions. The central claim is that the SMC is now applicable to gene flow, ecological life-history traits, structural variation, and full ancestral recombination graph inference, and the review is intended to make these developments accessible to ecologists and conservation biologists.","tokens_in":27876,"tokens_out":4669,"duration_ms":47656,"significance":"If its synthesis is accurate, this review provides a useful map of a rapidly moving field and could help non-specialist empiricists select appropriate SMC-based tools. Its strengths include a clear and mostly accurate primer (Section 1 and Algorithm 1), faithful descriptions of the cited methods, and a final section that explicitly acknowledges simulation-based validation limits and the need for users to check implementation choices. The paper is a review rather than a derivation, so the usual equation-level circularity checks do not apply; I found no internally inconsistent technical step. The one discrepancy identified by the stress-test concern, the Abstract's mention of 'inference of dispersal rates' without a supporting section, is real but local and editorial in nature, not a load-bearing flaw in the review's central synthesis.","major_comments":[],"minor_comments":[{"comment":"The Abstract and Introduction list 'inference of dispersal rates' among the recent advancements, but no section of the paper actually describes a method for inferring dispersal rates. The closest material is migration and gene-flow inference in Section 3.2 (MSMC-IM, MiSTI, cobraa). The authors should either add a brief discussion of dispersal-rate estimation or revise the list to avoid an unsupported claim.","section":"Abstract and Section 0"},{"comment":"Several highlighted methods are presented as ready-to-use tools without flagging that they are bioRxiv preprints rather than peer-reviewed publications: PSMC+ (Section 2.1), DoLoReS (Section 3.1), MiSTI (Section 3.2), SINGER (Section 4), and Threads (Section 4). Given the target audience of ecologists and conservation biologists, the manuscript should identify preprint status at first mention or include a column in a summary table indicating which tools have been peer-reviewed.","section":"Sections 2.1, 3.1, 3.2, and 4"},{"comment":"The manuscript does not include a conflict-of-interest statement even though several authors are developers of methods that receive prominent discussion (PSMC+, DoLoReS, cobraa). A competing-interests statement, or at minimum a sentence in the Acknowledgments, would make the review's positionality transparent.","section":"Acknowledgments"},{"comment":"Algorithm 1's step 4 states that the right emerging branch re-coalesces 'after a waiting time which is exponentially distributed, with rate proportional to the number of ancestral lineages at each epoch.' This is correct in spirit but could be clearer by specifying that the rate is the coalescent rate i/(2N_e) while the number of ancestral lineages i is constant within each epoch and changes only at coalescence events.","section":"Section 1.3 and Algorithm 1"},{"comment":"The caption describes type 1 events as 'not altering the next tree,' which is correct for SMC′ but could confuse readers who recall that the original SMC always produces a new marginal genealogy; a one-sentence reminder that the figure illustrates SMC′ rather than the original SMC would improve clarity.","section":"Figure 1 caption"}],"recommendation":"minor_revision","confidential_remarks":"The manuscript is a competent review and I expect the revisions to be straightforward. The main editorial concern is that several prominent recommendations rest on non-peer-reviewed preprints; this should be flagged clearly, even if the authors believe the methods are sound. I would also encourage the editor to ensure the journal's conflict-of-interest policy is satisfied, given the overlap between the author list and the highlighted tools."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"It's a solid review, not a research contribution, and the taxonomy is the only new thing in it. The three-part split (joint parameter inference, lineage partitioning, ARG inference) is a helpful organizing lens, and the SMC/SMC′ primer plus Algorithm 1 are accurate and clearly written. The method descriptions are mostly faithful to the primary literature, and the application examples are well chosen. Section 5 on data requirements, phasing, and the recency drawback is the most valuable part—it gives empiricists concrete practical warnings rather than just listing tools. The figures are informative, and they provide code for Figure 2B, which is more than most reviews do.\n\nSoft spots, in proportion. First, the abstract and introduction list \"inference of dispersal rates\" as an advance, but no section covers dispersal rates; the closest is migration and gene-flow inference in §3.2. That's an editorial mismatch, not a load-bearing flaw. Second, several recommended methods are from unreviewed preprints (DoLoReS, SINGER, Threads, MiSTI) without flagging them as such. For a review aimed at conservation biologists and ecologists who may not track preprint status, that's a real weakness—at minimum each such recommendation needs a caveat. Third, there is no conflict-of-interest statement even though several authors are developers of methods highlighted in the review (PSMC+, DoLoReS, cobraa). The treatment of those methods does not feel unfair, but the field has norms about disclosure and the omission is noticeable. None of these undercut the central synthesis: the claim that SMC-based methods now estimate parameters well beyond effective population size, and the cited methods do exist and do estimate those quantities. The paper is also honest about simulation-based validation limits in Section 6.\n\nWho is this for? A population geneticist who follows the field will not learn much, but an ecologist, conservation biologist, or graduate student wanting a map of the SMC landscape will get a genuinely useful orientation. It is a review, so it introduces no new result and its impact depends on how quickly the field moves.\n\nI would send this to peer review. It is accurate enough and useful enough to deserve referee time; the preprints and COI statement need to be addressed before publication, but those are fixable editorial issues.","headline":"A competent, genuinely useful review of SMC extensions beyond Ne, with a small editorial mismatch and some preprint-based recommendations that need flagging.","tokens_in":712,"tokens_out":1094,"would_cite":true,"duration_ms":26723,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The sequentially Markovian coalescent, once used mainly to estimate historical population sizes, has been extended to jointly infer recombination and mutation rate maps, gene flow, selfing and dormancy, structural variants, speciation…","keywords":["sequentially Markovian coalescent","demographic inference","gene flow","conservation biology","ARG reconstruction","recombination rate","structural variation","self-fertilization"],"falsifier":"A benchmark that applies the newer SMC extensions (for example, DoLoReS for structural variants, MiSTI for migration rates, eSMC for selfing) to genomes with independently known ground truth, such as experimental pedigrees, validated inversions, or historically documented admixture events, and finds systematic deviations from the truth would falsify the review's implied recommendation that these methods are ready for field adoption.","tokens_in":27547,"feed_emoji":"🧬","tokens_out":3510,"duration_ms":32049,"temperature":0.7,"pith_summary":"This review argues that the sequentially Markovian coalescent (SMC) has outgrown its original role as a tool for reconstructing historical effective population sizes ($N_e$). Recent extensions allow joint estimation of recombination and mutation rate variation, self-fertilization and germination rates, migration and admixture, structural variation, and full ancestral recombination graphs. The authors organize these methods into three classes: joint parameter inference, lineage partitioning, and ARG reconstruction. They make the case that ecologists, conservation biologists, and phylogeneticists should adopt these tools, while conceding that much of the validation is simulation-based and may not capture natural sequence complexity.","feed_headline":"Beyond $N_e$: SMC genomes now reveal gene flow, selfing, more","feed_subtitle":"The SMC framework now estimates recombination maps, migration rates, and mating systems, not just population size history.","key_machinery":"The central object is the SMC (more precisely the SMC′) approximation to the coalescent with recombination: a Markov process along the genome in which each local genealogy depends only on the genealogy at the previous position. This Markov property makes the model tractable for hidden Markov model (HMM) inference, with hidden states that are discrete coalescence-time bins and emissions that are the observed homozygous/heterozygous genotypes under an infinite-sites mutation model. A second key mechanism is lineage partitioning, where lineages are assigned to labeled classes (by inversion status, population, or species) and coalescence or recombination events are modulated between classes, enabling inference of migration, gene flow, and structural variation. The review also highlights threading-based ARG inference as a third mechanism that builds on the SMC to reconstruct full genealogical histories.","core_discovery":"The central claim is that the SMC model is a general and tractable probabilistic scaffold for the genealogical history of a sample, and that recent extensions have turned it into a multi-purpose inference engine. Where PSMC and MSMC read only $N_e$ from single or few genomes, methods such as iSMC, eSMC, MSMC-IM, MiSTI, cobraa, DoLoReS, and TRAILS now read recombination maps, mutation rate heterogeneity, mating-system parameters, migration rates, admixture times, structural variant locations, and speciation times from the same kind of sequence data. The review further claims that this broadening makes SMC-based inference relevant to conservation biology, natural history, and phylogenetic analysis, not just human population genetics.","pith_inferences":["The SMC transition matrix itself could serve as a summary statistic for approximate Bayesian computation to test competing structured demographic models, an idea the review only gestures at.","Extending SMC-based methods to accept genotype likelihoods and unphased low-coverage data would open the tools to ancient DNA and non-model organisms, which the review identifies as a desirable future direction.","The review implies a shift toward the 'ARG era' in which reconstructed genealogies become the primary data product; a testable consequence is that ARG-based estimators of selection and demography will gradually replace SFS-based approaches for many questions."],"forward_implications":["SMC-based methods can now identify genomic regions of conservation concern, such as low-recombining regions that are vulnerable to linked selection and runs of homozygosity.","Structural variants, including inversions that are invisible to read-based approaches, can be detected from single diploid genomes through localized suppression of recombination in inferred genealogies.","Migration and admixture times can be estimated from a single diploid sequence or from pairs of genomes, aiding studies of hybridization, introgression, and speciation.","Ecological life-history traits such as self-fertilization rate and seed-bank dormancy can be inferred jointly with demographic history, correcting biases in $N_e$ estimates.","Ancestral recombination graphs reconstructed under the SMC can be used to study selection, population size changes, and even speciation processes without fossil calibrations."],"supporting_citations":[{"why":"Introduces the SMC as a tractable Markovian approximation to the coalescent with recombination, the foundational model for all methods reviewed.","marker":"McVean and Cardin (2005)"},{"why":"Introduces SMC′, the more accurate variant that allows self-coalescence, which most modern SMC-based inference methods adopt.","marker":"Marjoram and Wall (2006)"},{"why":"Develops PSMC, the HMM that infers $N_e$ history from a single diploid genome and serves as the baseline for many extensions.","marker":"Li and Durbin (2011)"},{"why":"Develops MSMC, extending SMC inference to multiple genomes and introducing the cross-coalescence rate used for population separation analysis.","marker":"Schiffels and Durbin (2014)"},{"why":"Introduces MSMC2 and MSMC-IM, enabling inference of time-dependent migration rates under an isolation-with-migration model.","marker":"Wang et al. (2020)"},{"why":"Introduces the lineage-partitioning SMC variant for polymorphic inversions, the conceptual basis for detecting structural variation.","marker":"Peischl et al. (2013)"},{"why":"Derives the distribution of clade genomic span under the SMC and implements DoLoReS, which detects structural variants from reconstructed genealogies.","marker":"Ignatieva et al. (2025)"},{"why":"Develops eSMC, which jointly infers $N_e$ with self-fertilization or germination rates by exploiting their opposing effects on recombination and mutation.","marker":"Sellinger et al. (2020)"},{"why":"Extends eSMC to time-varying parameters (teSMC), allowing inference of transitions from outcrossing to selfing and associated demographic changes.","marker":"Strütt et al. (2023)"},{"why":"Develops ARGweaver, the threading-based Bayesian ARG inference method that relies on SMC dynamics and is a reference for later ARG tools.","marker":"Rasmussen et al. (2014)"}],"fun_headline_variants":["SMC beyond $N_e$: now infers gene flow, selfing, recombination","SMC expands: ecology to phylogenies, not just $N_e$","SMC: more than just $N_e$, now decodes ecology and phylogenies","From $N_e$ to everything: SMC now spans ecology, phylogeny","New SMC: from gene flow to selfing, beyond $N_e$"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The reviewed methods are assumed to be accurate enough outside simulation settings to guide real biological conclusions; the review itself warns that most are validated only against simulations that may not capture the complexity of natural sequences.","fun_headline_variants_meta":{"raw":{"variants":["SMC beyond $N_e$: now infers gene flow, selfing, recombination","SMC expands: ecology to phylogenies, not just $N_e$","SMC: more than just $N_e$, now decodes ecology and phylogenies","From $N_e$ to everything: SMC now spans ecology, phylogeny","New SMC: from gene flow to selfing, beyond $N_e$"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00087,"raw_usage":{"total_tokens":3742,"prompt_tokens":893,"completion_tokens":2849,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":509,"completion_tokens_details":{"reasoning_tokens":2741}},"tokens_in":509,"tokens_out":2849,"duration_ms":18447,"temperature":1.0,"reasoning_tokens":2741,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T12:00:01.598753+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A benchmark that applies the newer SMC extensions (for example, DoLoReS for structural variants, MiSTI for migration rates, eSMC for selfing) to genomes with independently known ground truth, such as experimental pedigrees, validated inversions, or historically documented admixture events, and finds systematic deviations from the truth would falsify the review's implied recommendation that these methods are ready for field adoption.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the SMC as a tractable Markovian approximation to the coalescent with recombination, the foundational model for all methods reviewed."},{"cited_title":"coalescent","cited_arxiv_id":null,"evidence_quote":"Introduces SMC′, the more accurate variant that allows self-coalescence, which most modern SMC-based inference methods adopt."},{"cited_title":"and Durbin, R","cited_arxiv_id":null,"evidence_quote":"Develops MSMC, extending SMC inference to multiple genomes and introducing the cross-coalescence rate used for population separation analysis."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces MSMC2 and MSMC-IM, enabling inference of time-dependent migration rates under an isolation-with-migration model."},{"cited_title":"F., and Kirkpatrick, M","cited_arxiv_id":null,"evidence_quote":"Introduces the lineage-partitioning SMC variant for polymorphic inversions, the conceptual basis for detecting structural variation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Derives the distribution of clade genomic span under the SMC and implements DoLoReS, which detects structural variants from reconstructed genealogies."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Develops eSMC, which jointly infers $N_e$ with self-fertilization or germination rates by exploiting their opposing effects on recombination and mutation."},{"cited_title":"D., Hubisz, M","cited_arxiv_id":null,"evidence_quote":"Develops ARGweaver, the threading-based Bayesian ARG inference method that relies on SMC dynamics and is a reference for later ARG tools."}],"review_version":1}