Fig. 1: Core claims of Monroe et al. are error artefacts.
From: Re-evaluating evidence for adaptive mutation rate variation
a, Top row: the profile of errors (misascribed as mutations) expected around TSS and TTS attributable to the errors associated with A/T homopolymeric runs. We simulated 2 million errors associated with A/T homopolymeric runs and used the Monroe script to generate the left plot. The data are an exact match to their somatic data calls (reproduced here as the top right plot, figure reused from Monroe et al.4 under Creative Commons Attribution 4.0 International License). Bottom row: germline mutations in TEs (left panel). Observed mutations in germline unmasked. bp, base pair. b, Essential genes do not have a low mutation rate. The second claim of Monroe to substantiate that mutation is lower in more functionally important sequences is that essential genes have the lowest mutation rates (their Fig. 3c). To substantiate this, they seem to have used many thousands of unfiltered calls. We repeat the analysis using filtered data, either Weng or Monroe, including indels. In neither case is there significant heterogeneity (we provide χ2 values for all comparisons, df = 3, but for Monroe CDS numbers are so small that these calculations are not valid and presented for completeness alone). Tests are one-sided in the sense that we call significance only if there is heterogeneity not if they are more similar than expected. Tests are two-sided in the sense that we ask about deviation from null in any direction. Unification of the two datasets does not alter conclusions: CDS, χ2 = 3.3; intron, χ2 = 5.76, P > 0.05 for all without multi-test correction. Error bars are s.e.m. across gene samples for which sample sizes are: essential (n = 719), morphological (n = 861), cellular or biochemical (n = 297) and environmental (n = 522) for CDS analysis (Monroe CDS and Weng CDS), and essential (n = 671), morphological (n = 789), cellular or biochemical (n = 270) and environmental (n = 452) for the intron analysis (Monroe intron and Weng intron). For representation of the underlying data, see Extended Data Fig. 2.
