Musa

LCN genes were identified based on OrthoFinder^{26 (link)} results. The orthologues were obtained from six monocots (Spirodela polyrhiza, Zostera marina, Musa acuminata, Ananas comosus, Sorghum bicolor and Oryza sativa) and six eudicots (Nelumbo nucifera, Vitis vinifera, Populus trichocarpa, A. thaliana, Solanum lycopersicum and Beta vulgaris), N. colorata, Amborella, and the gymnosperms G. biloba, P. abies and P. taeda. LCN genes needed to meet the following requirements: strictly single-copy in N. colorata, Amborella, G. biloba, P. abies or P. taeda, and single-copy in at least five of the 12 eudicots or monocots. With G. biloba, P. abies or P. taeda as the outgroup, we identified 2,169, 1,535 and 1,515 orthologous LCN genes, respectively. Furthermore, we trimmed the sites with less than 90% coverage. LCN gene trees were estimated from the remaining sites using RAxML v.7.7.8 using the GTR+G+I model for nucleotide sequences (Fig. 1c) and the JTT+G+I model for amino acid sequences (Supplementary Note 4.1). To account for incomplete lineage sorting and different substitution rates, we applied the multispecies coalescent model and a supermatrix method, respectively, to the LCN genes and found further support for the sister relationship between Amborella and all other extant flowering plants (Supplementary Note 4.2).
We further carefully selected five LCN gene sets (1,167, 834, 683, 602 and 445) from 115 species and applied both a supermatrix method^{27 (link)–29 (link)} and the multi-species coalescent model to infer the phylogeny of angiosperms (Supplementary Note 4.2). The phylogeny inferred from 1,167 LCN genes is shown in Fig. 1d, with different support values from the multi-species coalescent analyses of the other four LCN gene sets.
To estimate the evolutionary timescale of angiosperms, we calibrated a relaxed molecular clock using 21 fossil-based age constraints^{7 (link)} throughout the tree, including the earliest fossil tricoplate pollen (approximately 125 Ma) associated with eudicots³⁰ . We concatenated 101 selected genes (205,185 sites) and fixed the tree topology to that inferred from our coalescent-based analysis of 1,167 genes from 115 taxa. We performed a Bayesian phylogenomic dating analysis of the 101 selected genes in MCMCtree, part of the PAML package^{31 (link),32 (link)}, and used approximate likelihood calculation for the branch lengths^{33 (link)}. Molecular dating was performed using an auto-correlated model of among-lineage rate variation, the GTR substitution model, and a uniform prior on the relative node times. Posterior distributions of node ages were estimated using Markov chain Monte Carlo sampling, with samples drawn every 250 steps over 10 million steps following a burn-in of 500,000 steps. We checked for convergence by running the analysis in duplicate and checked for sufficient sampling.
We also implemented the penalized likelihood method under a variable substitution rate using TreePL^{34 (link)} and r8s^{35 (link)}, as a constant substitution rate across the phylogenetic tree was rejected (P < 0.01) for all cases by likelihood-ratio tests in PAUP³⁶ . Three fossil calibrations, corresponding to the crown groups of Lamiales, Cornales and Laurales, were implemented as minimum age constraints in our penalized likelihood dating analysis, except that the earliest appearance of tricolpate pollen grains (about 125 Ma)³⁰ was used to fix the age of crown eudicots. We determined the best smoothing parameter value of the concatenated 101 LCN genes as 0.32 by performing cross-validations of a range of smooth parameters from 0.01 to 10,000 (algorithm = TN; crossv = yes; cvstart = −2; cvinc = 0.5; cvnum = 15). We used 100 bootstrap trees with branch lengths generated by RAxML^{37 (link)} to infer the 95% confidence intervals of age estimates (Supplementary Note 4.2).

The OrthoFinder^{97 (link)} clustering method was used to classify complete proteomes of 23 sequenced green plant genomes, including A. filiculoides and S. cucullata (Supplementary Table 5), into orthologous gene lineages (that is, orthogroups). We selected taxa that represented all of the major land plant and green algal lineages, including six core eudicots (A. thaliana, Lotus japonicus, Populus trichocarpa, Solanum lycopersicum, Erythranthe guttata and Vitis vinifera), five monocots (O. sativa, Sorghum bicolor, Musa acuminata, Zostera marina and Spirodella polyrhiza), one basal angiosperm (A. trichopoda), two gymnosperms (Pinus taeda and Picea abies), two ferns (A. filiculoides and S. cucullata), one lycophyte (S. moellendorffii), four bryophytes (Sphagnum fallax, P. patens, Marchantia polymorpha and Jungermannia infusca) and two green algae (Klebsormidium flaccidum and C. reinhardtii). In total, 16,817 orthogroups containing at least two genes were circumscribed, 8,680 of which contain at least one gene from either A. filiculoides or S. cucullata. Of the 20,203 annotated A. filiculoides genes and the 19,780 annotated S. cucullata genes, 17,941 (89%) and 16,807 (84%) were classified into orthogroups, respectively. The details for each orthogroup, including gene counts, secondary clustering of orthogroups (that is, super-orthogroups)^{110 (link)} and functional annotations, are reported in Supplementary Table 5.
We used Wagner parsimony implemented in the program Count^{111 (link)} with a weighted gene gain penalty of 1.2 to reconstruct the ancestral gene content at key nodes in the phylogeny of the 23 land plants and green algae species (Supplementary Table 5). The ancestral gene content dynamics—gains, losses, expansions and contractions—are depicted in Supplementary Fig. 5. Complete details of orthogroup dynamics for the key ancestral nodes that include seed plants, such as Salviniaceae, euphyllophytes and vascular plants, are reported in Supplementary Table 5.

Two gene annotations of the Musa acuminata draft genome sequence were available for the initial assembly. The first corresponded to the annotation published by [27 (link)], in addition to approximately 1000 genes curated by human expertise before 08 December 2014 (http://banana-genome.cirad.fr/). The second one was the NCBI RefSeq genome annotation released the 7 October 2014 (ftp://ftp.ncbi.nlm.nih.gov/genomes/refseq/plant/Musa_acuminata/) and generated with the NCBI Eukaryotic Genome Annotation Pipeline.

A purposive method of sampling was used. Using this technique, the researcher was able to select the students who had relevant experiences to the topic being researched (Etikan, Musa & Alkassim 2016 (link)). At the time of the study, the population included N = 1001 undergraduate nursing students registered in the 4-year undergraduate programme at the selected School of Nursing (Donough et al. 2014 ). The coordinators at the institution under study helped with the selection of the undergraduate nursing students who met the inclusion criteria. This was performed after ethical clearance when the Health Research Ethical Committee at Stellenbosch University and the head of the department at institution under study gave approval.
The focus groups included nine registered undergraduate students per year level of the Bachelor’s degree in Nursing as part of the inclusion criteria. Sequentially, there were nine first-year students, nine second-year students, nine third-year students and nine fourth-year students who took part in the group interviews. Furthermore, the inclusion criteria required that the students to have worked at a clinical placement setting where they received clinical supervision. Thirty-six students from nursing formed the sample. According to Creswell and Poth (2016 ), in qualitative research, the number of participants is sufficient once information in the particular subject area has reached saturation and had been verified. In addition, instead of the number of participants, information that is of quality is the main focus of qualitative research.

In an attempt to investigate possible relations to reading habits or to musical proficiency, further data was collected. Reading habit questionnaire: A questionnaire was developed to measure reading habits and their potential correlation with the overall reading performance of individuals. The following data were collected: Demographic data (age, gender, educational level), reading habits I (categories, e.g., newspaper, novels, etc.), reading habits II (percentage of reading/writing, analog vs. digital, and reading time spend with category), reading habits III (actual familial reading habits and during childhood, reading to other people privately/professionally), speaking/writing development, speaking habits, L2 languages, language (therapy) experiences, speaking anomalies (e.g., mumbling), and auditive habits (volume setting tendencies, i.e., loud/silent, music listening preferences). MusA. A short Questionnaire to Assess Musical Activity (Fernholz et al., 2018 (link)), which investigates music preferences as well as musical activity. After analyzing the two questionnaires for possible correlations between musicality and reading habits first, only factor musical was integrated in later analysis.

Mature green banana fruit (Musa acuminata AAA group, Cavendish subgroup) was obtained from a local market in Guangdong, China. Ethephon-induced ripening, 1-MCP-delayed ripening, and natural ripening of the fruit were evaluated as described previously (Li et al., 2020 (link)). Each treatment contained three biological replicates and was stored at 22°C with about 90% relative humidity until fully ripe. The fruit samples from various development stages were collected from local banana plantations in Guangzhou, China. The sampling time point was determined by the number of days after flowering. The fruit samples at each time point were collected and quickly frozen in liquid nitrogen and then stored at −80°C until utilization.