simSOMA: a cell-lineage based simulator of the somatic VAF spectrum in plants

Key results

1. SuggestedIn a factorial scan on a 20-organ branched topology, smaller niches, higher turnover, fewer branch or organ precursors and greater developmental depth all raised the fixed fraction and lowered the intermediate fraction, so these parameters act in overlapping ways.

   Grid of m from 2 to 20, 16 ρ values between 0 and 1, 12 Pb values (1 to 20) and 12 Po values (1 to 30), with 50 replicates per combination; C = 64, organs of 10^4 cells with 1% (100 cells) sampled, organs on branch orders 1 to 5. Organs at developmental depth 6 with small m gave bimodal spectra; depth 2 with larger m kept more unfixed variants at or above τ = 0.05. · Global parameter scan identifies key VAF…

2. SuggestedTerminal organ growth produced most private variants, concentrated at low VAF, with the cumulative count above frequency q linear in 1/q; the slope reflects the per-division mutation rate during organ growth, while Po mainly shaped variants already present at organ initiation.

   m = 4, ρ = 0, C = 64 and Pb = 4 held fixed on the 20-organ topology while Po and O varied. Growing from Po to O cells takes O − Po divisions (2(O − Po) daughter cells); about 2(1/q − Po) daughters seed at least a fraction q of the organ, a prediction that agreed closely with full simulations (Supplementary Fig. S3). Private fraction rose strongly with O and stayed high across Po for large O. · Organ growth is the main source of private…

3. ShownWithin the model, the chance that a branch is founded by more than one clonal sector depends only on ϕ = (Peff − 1)m/C, with Peff = min(Pb, m) (for C ≥ m), so branch founding is predominantly monoclonal unless C < m(m − 1), and Pb cannot be identified from mutation data beyond m.

   Topology with five lateral branches, one sampled organ each; fixed fraction scanned over Pb and C at m = 3 and m = 12 fell with Pb and plateaued once Pb reached m. With a uniform focal position and C divisible by m, π ≤ (m − 1)m/C, so monoclonal founding predominates unless C < m(m − 1); organ founding, where Po is not capped by m, can be strongly polyclonal. · Branching bottlenecks depend on precursor number…; Conclusion and Discussion; Figure legends

4. SuggestedLayer mixing and unphased readout rescale layer-fixed variants to lower VAFs, and read-level sampling mainly removes the low-frequency and minor-layer signal: in simulated unphased bulk data at mean depth 60 only the L2-fixed peak stayed clearly resolvable, and most private organ-growth variants go undetected.

   Ten simulated trees, three independent layers with contributions (0.10, 0.70, 0.20); the exact unphased bulk transformation places layer-fixed variants at 0.05, 0.35 and 0.10. At mean depth 60, L1- and L3-fixed variants were still 78% and 93% detected but overlapped low-frequency variants. True VAF at which half of variant-organ occurrences were detected: about 0.09, 0.03 and 0.013 at mean depths 20, 60 and 150. · From developmental to observed VAF spectra; Thus, for a three-layer bulk sample with…; Read-level observation; Figure legends

5. SuggestedOn a single 200-year axis, niche size m and turnover ρ dominated the fixed, intermediate, private and sharedness statistics; higher ρ raised both private fraction and sharedness, while spiral phyllotaxy left these marginals nearly unchanged but added structure to pairwise organ sharing.

   20 organs sampled at regular positions; smaller m gave higher fixed fractions and higher ρ raised fixation for most m. With spiral phyllotaxy, off-diagonal pairwise Jaccard sharing appeared at ρ = 0 and was strongly reduced at ρ = 1. Values of m up to 20 were simulated, but the paper treats m ≤ 4 as the typical range (2 to 4 persistent ASCs per layer). · Stem-cell turnover shapes VAF spectra and…; Figure legends