Chapter 6 Mutations detection
This chapter briefly introduced the next chapters on the detection of mutations:
- Using an in silico virtual experiment
- Using swiss data from Schmid-Siegert et al. (2017)
- Using french data from Plomion et al. (2018)
- Using Angela sequences
- Using sequences from the Sextonia rubra
- Using sequences from the “Faux de Verzy”
The three first analyses resulted in a published manuscript: https://peercommunityjournal.org/articles/10.24072/pcjournal.187/
6.1 In silico experiment
The in silico experiment uses the generateMutations and the detectMutations workflows to test the effect of coverage/sequencing depth and allelic frequency on the performance of 7 tools to detect mutations either generalist or specific to mutations.
Figure 6.1: Generate mutations workflow.
Figure 6.2: Detect mutations workflow.
6.2 Swiss data - Schmid-Siegert et al. (2017)
We re-analyzed mutations from Schmid-Siegert et al. (2017) with Strelka2 and GATK to compare original and obtained mutations to show the interest of Strelka2 and high sequencing depth.
Figure 6.3: Detect mutations workflow for Swiss data.
6.3 Bordeaux data - Plomion et al. (2018)
We re-analyzed mutations from Plomion et al. (2018) with Strelka2 and Mutect2 to compare original and obtained mutations to show the interest of Strelka2 and high sequencing depth.
Figure 6.4: Detect mutations workflow for Swiss data.
6.4 Angela
Genome, heterozygous sites, cambial mutations and leaf mutations from sequences of Angela.
Figure 6.5: Detect mutations workflow for Angela.
6.5 Sextonia rubra
Genome, heterozygous sites, cambial mutations and leaf mutations from sequences of Sixto
Figure 6.6: Detect mutations workflow for Sixto.
6.6 Hetre
Mutations from sequences of the “Faux de Verzy”.
Figure 6.7: Detect mutations workflow for Angela.
6.7 Project description
For each species, we will extract DNA and construct 27 individually tagged genomic libraries to produce short read sequencing (see sampling strategy, WP2 above) with NovaSeq6000 SP technology, targeting aminimum of 50x coverage per library (70-90x should be realistic, we will go for 100x if possible). The possible sequencing depth affordable in the project will depend on the genome size of chosen species and some extent on fluctuating sequencing prices. Lab work will be conducted at Ecofog and PGTB (Biogeco); the sequencing will be subcontracted to CEA Genoscope, Evry.
Detection of mutations (T4.2): For mutation detection, we will minimise the impact of library preparation and sequencing errors by retaining only mutations detected by comparison with the “zero mutation reference” (Hanlon et al 2019) and in each branch (Orr et al. 2019), giving a total of 9 genotype sampling points in each tree. We will quantify the occurrence of mutations in each of the 24 leaf libraries per tree using a high-sensitivity method (Cibulskis et al., 2013). De novomutations appear in single meristem cells and their abundance in tree tissues will depend on the specific cell’s divisions and contribution to generating tree tissues. Fixation of de novo mutations in tissues is expected to be rare and the expected time to fixation is long, see (Nicholson et al., 2018)for an example in humans. The detection of novel mutations must thus be sensitive to low allele copy number in tissues. For this reason, we will test and use mutation detection methods typically used in cancer mutation research, that are optimized for strong allelic frequency skews (Alioto et al., 2015).The artificial inclusion of in silicomutationsinto the reads and permutations among branch labels allows the estimation of the recovery rate of true mutations, the false negative rates, and the false positive rates (Orr et al., 2019). We will test whether the rate of accumulation of mutations is increased in high-light-exposed branches compared to those with low-light exposure accounting for the physical tree branches (tree architecture -WP3), the mutation history, and light environment (WP3).