WES vs ONT DNM allele frequency comparison and parent-of-origin
In total, 77 DNMs of 109 DNMs (71%) were phased and the
parent-of-origin determined, with 64 of these (83%) being of paternal
origin. Across the 77 phased DNMs there was a DNM allele frequency range
of 6% to 78% in the ONT data and 15% to 71 % in the WES DNM base
data, clearly showing the presence of postzygotic and prezygotic
mutations. From these 77 DNMs, our targeted long-read sequencing
approach identified 61 clearly prezygotic DNMs, with a possible third
allelic form found in the remaining 16 DNMs. The separation of pre and
postzygotic DNMs can be established more clearly through a combined
analysis of their base frequencies, allele frequencies, DNM wt allele
frequency, and background sequencing error (see Supplementary Table 11).
By doing this we confirmed 69 prezygotic DNMs and 8 postzygotic
mutations, as depicted in Supplementary Table 11 and 12, and
Supplementary Figure 4 and 5. On average a third of allelic data is
identified as false in our ONT approach (Supplementary Table 7), and
that impacted the allele frequency significantly. After removal of these
false allelic data, we noticed that the average ONT allele frequency for
prezygotic DNMs were around 50% with an average of 49.6%, with a
standard error margin (SEM) of 0.8% (Supplementary Table 13 and
Supplementary Figure 4). For germline DNMs, we expect the mutant allele
frequency to be approximately 50%. shows that the majority of all ONT
DNM allele frequencies were around 50% with the prezygotic average of
49.6% (SEM 0.8%). While raw allele frequencies were expectedly much
worse than DNM base frequencies, the polished allele frequencies are
slightly more accurate compared to the WES DNM and ONT DNM prezygotic
nucleotide base frequencies of 48.6% (SEM 1.0%) and 48.8% (SEM
0.8%), respectively (Supplementary Table 13). As expected, postzygotic
frequencies of mutant alleles deviated significantly from this, with an
average of 15.9% and a standard error margin of 2.4% in ONT allele
data, though the frequencies of postzygotic mutations is less consistent
in WES data, where the postzygotic frequency average was 28.8% (SEM
4.9%) (Supplementary Table 13).
The parent-of-origin was also investigated within the pre and
postzygotic fractions, where we observe the prezygotic paternal
preference of 83% and the postzygotic paternal origin of 62% (5 out of
8), see Supplementary Figure 5. Classification of postzygotic DNMs was
determined as per Supplementary Table 11 and a detailed IGV illustration
is depicted in Figure 3. All postzygotic calls were corroborated by the
DNM base frequencies in the short-read and long-read data, and further
supported by all identified allele frequencies and the degree of error
in target data (see Supplementary Table 7 and 11).