Pedigree and litter size
Except for pig 7, all the RT carriers studied here had relatives reported to carry the same RT, suggesting they were not de novo but inherited from a parent. The expectation is that RT carriers have reduced number of live born litter sizes because half of the gametes they produce are unbalanced, leading to unbalanced foetuses, which are most likely unviable. Therefore, we had a look at the litter size of the RT carriers and their sires and dams (Table 1). The RT carriers belong to a boar line with an average litter size of 10 piglets. In general, the sires showed a higher average litter size than the dams, because for the sires also crossbred litters were included. In addition, the sire averages were based on much larger number of litters and hence showed lower standard deviations.
Table 1
Litter size and number of litters with records for the studied individuals and their parents
| Individual | Sire | Dam |
| Litter size1 | N2 | Litter size | N | Litter size | N |
PIG1 | - | 0 | 12.2 (± 3.7) | 174 | 10.4 (± 3.3) | 5 |
PIG2 | - | 0 | 14.0 (± 3.9) | 275 | 8 (± 7.1) | 2 |
PIG3 | - | 0 | 14.4 (± 3.9) | 211 | 10.5 (± 2.4) | 6 |
PIG4 | - | 0 | 14.4 (± 3.9) | 211 | 10.5 (± 2.4) | 6 |
PIG5 | - | 0 | 6.6 (± 2.9) | 13 | 9.4 (± 1.6) | 7 |
PIG6 | - | 0 | 12.0 (4.0) | 55 | 10.5 (± 2.4) | 6 |
PIG7 | 5 (± 2.6) | 6 | 6.5 (± 2.4) | 6 | 8.5 (± 1.7) | 4 |
1 Litter size based on liveborn piglets only |
2 N is the number of litters with recorded litter size |
Table 1 shows that pig 7 sired litters himself (29 in total, but only 6 with litter size records) and showed a reduced average live born litter size of 5 (± 2.6). Given an average litter size of 10 for this boar line, this suggests that indeed half of the litter was unviable. Also the sires of Pig 5 and 7 had a reduced litter size, suggesting these sires are carriers of the respective RT and the RT was inherited rather than de novo. The dam of pig 2 had and average litter size of 8 with a large standard deviation. She gave birth to only 2 litters, one with 13 liveborn and 2 stillborn piglets, and the second with only 3 liveborn piglets of which pig 2 was one. The sire of pig 2 showed average litter size (14.0 ± 3.9) based on a large number of litters (275), suggesting the RT was inherited from the dam. Which can be supported by the fact that she gave birth to another known RT carrier in her first litter (i.e. maternal half sib of pig 2).
Pig 1 was a maternal half sib of the dam of pig 3, 4, and 6. Pig 3 and 4 were littermates, and pig 6 was a maternal half sib of them, see pedigree graph in Fig. 3. They all carried the same RT, t(2;4), which they most likely inherited from their (untested) (grand)mother. Surprisingly, for pig 1, 3, 4, and 6 there was no clear reduction in litter size from their parents. Given their pedigree relation, the dam of pig 3, 4, and 6, and dam of pig1 (maternal granddam of pig 3, 4, and 6) are most likely carriers of t(2, 4) (Fig. 3). However, these dams had an average live born litter size of 10.5 (± 2.4) and 10.4 (± 3.3), respectively, and none of the sires involved showed a reduced litter size.
These litter sizes and pedigree relations suggest that all studied RT carriers inherited the RT from a parent rather than a de novo occurrence of the RT.
Blind detection of RT in short read sequence data
Table 2 shows the resulting number of inter-chromosomal translocations from each step in our analysis. After visual inspection, we correctly discovered 6 out of the 7 RTs in a blind analysis (JBrowse images of the aligned reads on chromosome regions involved are given in Supplementary Fig. 1). The t(1;16) from pig7 was not detected in the blind analysis. All final detected RTs involved the chromosomes that match the RT results from karyotype staining. Therefore, we are confident these are the actual breakpoints causing the RT. For the related pigs with t(2;4), we detected the exact same translocation positions in all 4 pigs.
Table 2
Translocation output at various stages of the pipeline for 7 carriers
| PIG1 | PIG2 | PIG3 | PIG4 | PIG5 | PIG6 | PIG7 |
Karyotype | t(2;4) | t(6;8) | t(2;4) | t(2;4) | t(7;14) | t(2;4) | t(1;16) |
Coverage | 32.6 | 37.9 | 33.4 | 35.1 | 37.2 | 31.1 | 30.3 |
DELLY out | 73,923 | 96,378 | 85,597 | 94,223 | 94,683 | 75,368 | 78,946 |
Basic filt | 1,127 | 1,426 | 1,185 | 1,175 | 1,503 | 1,067 | 1,116 |
Final filt1 | 30(15) | 56(28) | 34(17) | 38(19) | 68(34) | 38(19) | 44(22) |
Visual insp | 1(2,4) | 1(6,8) | 1(2,4) | 1(2,4) | 1(7,14) | 1(2,4) | 0(FN2) |
1 one of the filter criteria for the detected inter-chromosomal translocations was the presence of a matching pair, hence between brackets are the number of reciprocal pairs, i.e. the actual number of possible RTs |
2 FN = false negative |
The 15 non-carrier animals all came out negative using the same filtering criteria and visual inspection as the carriers. Table 3 provides a summary of the results from the pipeline for the non-carriers.
Table 3
Translocation output at various stages of the pipeline for 15 non-carriers with negative RT results
| Average | Min | Max |
Coverage | 30.3 | 28.8 | 34.3 |
DELLY out | 57,748.6 | 41,341 | 70,218 |
Final filt | 35.7 | 16(8) | 58(29) |
Visual insp | 0 | 0 | 0 |
Analysis of breakpoints and junctions
The sequence data made it possible to refine the RT location and to investigate the breakpoints and junctions of the three detected RTs. For the t(2;4) translocation, we had four related individual showing the exact same breakpoints and junctions at 2:4983988/ 4:81209353 and 2:4983990/ 4:81209358, confirming the RT was inherited. Both breakpoints were located within a gene. The break on chromosome 2 was located in an intron of ENSSSCG00000032003, which is an ortholog of the human BRCA2 gene, a breast cancer gene involved in DNA repair. The break on chromosome 4 was located in an exon of C1orf112, which is an uncharacterized open reading frame.
For t(2;4), the breakends connected 5’ to 5’ (head of chr2 to head of chr4) and 3’ to 3’ (tail of chr2 to tail of chr4), see Fig. 4A for schematic representation. At the 3’ to 3’ junction there was a blunt end ligation. The 5’ to 5’ junction showed microhomology, i.e. there was an overlap of 2 bp in the sequence of both breakends, at 2:4983987–4983988 and 4:81209355 − 81209354 (Fig. 5A). From chromosome 2 one base was lost (2:4983989). From chromosome 4, two bases were lost (4:81209356–81209357). A normal copy of chromosome 2 has a length of 151.9 Mb, while a normal copy of chromosome 4 is 130.9 Mb long. The 3’ to 3’ junction resulted in a chromosome length of 196.7 Mb (147.0 Mb from chromosome 2 and 49.7 Mb from chromosome 4), matching the larger derived chromosome 2 in the karyotype picture (Fig. 1A). The 5’ to 5’ junction resulted in a chromosome length of 86.2 Mb (5.0 Mb from chromosome 2 and 81.2 Mb from chromosome 4), matching the smaller derived chromosome 4 in the karyotype picture (Fig. 1A).
We mapped the region of the t(6;8) translocation to 6:27901326/ 8:1266615 and 6:27901330/ 8:1266598. Both breakpoints were located within a gene. The breakpoint on chromosome 6 was located in an intron of SLC9A5, which is involved in pH regulation to eliminate acids generated by active metabolism or to counter adverse environmental conditions. The breakpoint on chromosome 8 was located in an intron of ZFYVE28, which is a negative regulator of epidermal growth factor receptor signalling.
For t(6;8), the breakends connected 5’ to 3’ (head of chr6 to tail of chr8) and 3’ to 5’ (tail of chr6 to head of chr8), see Fig. 4B for schematic representation. At the 3’ to 5’ junction, there was a blunt end ligation. The 5’ to 3’ breakpoint junction showed microhomology of 3 bp at 6:27901327–27901329 and 8:1266615–1266617 (Fig. 5B). For chromosome 8, fifteen bases were lost (8:1266599–1266614). A normal copy of chromosome 6 has a length of 170.8 Mb, while a normal copy of chromosome 8 is 139.0 Mb long. The 5’ to 3’ junction resulted in a chromosome length of 140.5 Mb (2.8 Mb from chromosome 6 and 137.7 Mb from chromosome 8), matching the slightly larger derived chromosome 8 in the karyotype picture (Fig. 1B). The 3’ to 5’ junction resulted in a chromosome length of 169.3 Mb (168.0 Mb from chromosome 6 and 1.3 Mb from chromosome 8), matching the slightly shorter derived chromosome 6 in the karyotype picture (Fig. 1B).
We mapped the region of the t(7;14) translocation to 7:118889969/ 14:49733352 and 7:118889973/ 14:49733364. The breakpoint on chromosome 7 was intergenic, while the one on chromosome 14 was located in an intron of the gene CABIN1.
For t(7;14), the breakends connected 5’ to 3’ (head of chr7 to tail of chr14) and 3’ to 5’ (tail of chr7 to head of chr14), for schematic representation see Fig. 4C. At the 3’ to 5’ junction, there was a blunt end ligation. In the 5’ to 3’ junction a micro-insertion of 12 novel bases was observed (Fig. 5C). For chromosome 7, two bases were lost, being 7:118889971–118889972. A normal copy of chromosome 7 has a length of 121.8 Mb, while a normal copy of chromosome 14 is 141.8 Mb long. The 5’ to 3’ junction resulted in a chromosome length of 210.9 Mb (118.9 Mb from chromosome 7 and 92.0 Mb from chromosome 14), matching the much larger derived chromosome 7 in the karyotype picture (Fig. 1C). The 3’ to 5’ junction resulted in a chromosome length of 52.7 Mb (3.0 Mb from chromosome 7 and 49.7 Mb from chromosome 14), matching the much shorter derived chromosome 14 in the karyotype picture (Fig. 1C).
False negative detection of t(1;16)
The RT t(1;16) for pig 7 was not detected in the blind analysis, i.e. false negative (Table 2). After filtering, there was one potential pair of translocations involving chromosome 1 and 16 (1:241933142/ 16:75687308, 1:241932872/ 16:75687319). However, it did not pass the visual inspection, as the breakpoint on chromosome 1 had overlapping forward and reverse reads (like Fig. 2C). RepeatMasker showed that there was a porcine repetitive SINE element of the PRE-1 family (22) on both involved chromosomes at the breakpoint locations (1:241932894–241933143 and 16:75687060–75687314), causing the discordant pairs and split alignments. In addition, these inter-chromosomal translocations showed up in the sequences of several other pigs indicating it is a common rearrangement due to the repetitive element and not the RT.
After unblinding, we investigated all translocations involving chromosome 1 and 16 from the raw DELLY output, because the RT may have been reported by DELLY, but may not have fulfilled all filtering criteria. DELLY detected 375 inter-chromosomal translocations between chromosome 1 and 16. Among those, we selected the 56 translocations with a matching reciprocal translocation with matching connection type (3’ to 3’ and 5’ to 5’ or 3’ to 5’ and 5’ to 3’), resulting in 28 possible RT pairs. For only two of those pairs, both breakends on the same chromosome were within 100 bp from each other. One had overlapping forward and reversed reads (like Fig. 2C) on chromosome 1, and was present in multiple animals, hence unlikely to be the true RT. While the other showed good characteristics of an RT on chromosome 1 (1:97320432–97320434; Fig. 6A), and a potential RT breakpoint in a repetitive region on chromosome 16 (16:23532186–23532204; Fig. 6B). Hence we investigated this pair further.
Based on the DELLY output, one translocation (1:97320434/ 16:23532200) was mapped with high quality and passed de criteria, but the other translocation (1:97320432/ 16:23532204) had a low mapping quality of 12, only 3 discordant read pairs and no split alignments to support the junction. Looking at it visually (Fig. 6), there was actually sufficient support for an inter-chromosomal translocation at 1:97320431 and 16:23522204, supported by split alignments partially mapping to each location. However, most forward reads were discordantly paired with a mate mapping to many other locations on the genome due to a Pre0_SS element on chromosome 16 at 23532226–23532486. This repetitive element led to the low quality and imprecise labels in the DELLY output. This background due to the repetitive element of Pre0_SS was also observed in other pigs at this location on chromosome 16, but without the split alignments supporting the RT for Pig7. This indicates that the blind analysis of DELLY output in combination with the filtering criteria were not suitable to detect this RT located in a repetitive region.
We mapped the region of the t(1;16) translocation to 1:97320431/ 16:23532204 and 1:97320436/ 16:23532186. The breakpoint on chromosome 1 was intergenic, while the one on chromosome 16 was located in an intron of the gene EGFLAM.
For t(1;16), the breaks connected 5’ to 3’ (head of chr1 to tail of chr16) and 3’ to 5’ (tail of chr1 to head of chr16), for schematic representation see Fig. 6C. At both junctions, there were micro-insertions of 15 and 20 novel bases (Fig. 6D). For chromosome 1, four bases were lost being 1:97320432–97320435. For chromosome 16 seventeen bases were lost, being 16:23532187–23532203. A normal copy of chromosome 1 has a length of 274.3 Mb while a normal copy of chromosome 16 is 79.9 Mb long. The 5’ to 3’ junction resulted in a chromosome length of 153.7 Mb (97.3 Mb from chromosome 1 and 56.4 Mb from chromosome 16), matching the larger derived chromosome 16 in the karyotype picture (Fig. 1D). The 3’ to 5’ junction resulted in a chromosome length of 200.5 Mb (23.5 Mb from chromosome 1 and 177.0 Mb from chromosome 16), matching the shorter derived chromosome 1 in the karyotype picture (Fig. 1D).
Reducing sequencing coverage
Randomly reducing the sequencing coverage of the carriers to 10, 15, and 20 fold coverage had a big impact on the blind detection of the RTs (Table 4). With 10 fold coverage not all the actual RT breakpoints were detected by DELLY (pig 1, 2, 6), and many of the ones that were detected had a low quality label (pig 3, 4, 5). Only the RT of pig 5 could be detected with relaxed filtering criteria. This indicates that 10 fold coverage is not enough for proper detection of RT.
Table 4
Results of reduced sequencing depth, with both strict and relaxed filtering criteria
Animal | | PIG1 | PIG2 | PIG3 | PIG4 | PIG5 | PIG6 | PIG7 |
Karyotype | t(2;4) | t(6;8) | t(2;4) | t(2;4) | t(7;14) | t(2;4) | t(1;16) |
10 fold | Coverage | 10.2 | 11.9 | 10.5 | 11.0 | 11.7 | 9.8 | 9.5 |
Strict1 | DELLY out | 18,772 | 26,448 | 22,308 | 25,634 | 18,809 | 14,042 | 20,364 |
filtering | Basic filt | 146 | 267 | 166 | 168 | 209 | 139 | 146 |
| Final filt3 | 4(2) | 6(3) | 0 | 2(1) | 4(2) | 0 | 4(2) |
| Visual insp | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
10 fold | Basic filt | 2739 | 3939 | 3007 | 3084 | 3675 | 2710 | 2831 |
Relaxed2 | Final filt3 | 32(16) | 24(12) | 28(14) | 26(13) | 40(20) | 8(4) | 28(14) |
filtering | Visual insp | 0 | 0 | 0 | 0 | 1(7,14) | 0 | 0 |
15 fold | Coverage | 16.8 | 19.5 | 17.2 | 18.1 | 19.2 | 16.1 | 15.6 |
Strict1 | DELLY out | 34556 | 46817 | 40754 | 45934 | 44923 | 35662 | 37186 |
filtering | Basic filt | 405 | 598 | 463 | 479 | 548 | 423 | 429 |
| Final filt3 | 8(4) | 16(8) | 12(6) | 18(9) | 24(12) | 4(2) | 14(7) |
| Visual insp | 0 | 0 | 0 | 0 | 1(7,14) | 0 | 0 |
15 fold | Basic filt | 5854 | 7769 | 6316 | 6537 | 7465 | 5738 | 6009 |
Relaxed2 | Final filt3 | 46(23) | 62(31) | 46(23) | 54(27) | 60(30) | 44(22) | 52(26) |
filtering | Visual insp | 0 | 1(6,8) | 1(2,4) | 1(2,4) | 1(7,14) | 1(2,4) | 0 |
20 fold | Coverage | 20 | 23.3 | 20.5 | 21.6 | 22.9 | 19.2 | 18.6 |
Strict1 | DELLY out | 33034 | 46672 | 39783 | 45198 | 44767 | 34398 | 35871 |
filtering | Basic filt | 569 | 697 | 549 | 621 | 717 | 554 | 564 |
| Final filt3 | 20(10) | 28(14) | 16(8) | 26(13) | 30(15) | 6(3) | 30(15) |
| Visual insp | 0 | 0 | 0 | 0 | 1(7,14) | 0 | 0 |
20 fold | Basic filt | 7,323 | 9,586 | 7,869 | 8,215 | 9,505 | 7,210 | 7,477 |
Relaxed2 | Final filt3 | 54(27) | 68(34) | 48(24) | 66(33) | 62(31) | 50(25) | 54(27) |
filtering | Visual insp | 1(2,4) | 1(6,8) | 1(2,4) | 1(2,4) | 1(7,14) | 1(2,4) | 0 |
1 Strict filtering: Basic filtering: Pass DELLY quality filter and precise position; final filtering: mapping quality = 60, at least 10 split reads supporting the translocation, ≤ 60 discordant paired end reads, consensus alignment quality > 0.9, and matching second translocation with matching connection type present |
2 Relaxed filtering: Translocations that failed the quality filter of DELLY (LowQual instead of PASS; basic filtering) and with 5 or more split reads supporting the translocation (final filtering) were retained. |
3 one of the filter criteria for the detected inter-chromosomal translocations was the presence of a matching pair, hence between brackets are the number of reciprocal pairs, i.e. the actual number of possible RTs |
With 15 fold coverage only the RT for pig 5 was detected, however for 4 other carriers both breakpoints were initially detected by DELLY but lost due to filtering. The filtering criteria for samples with 30 fold coverage can be set stringent to remove as many false positive as possible, for lower fold coverage these filtering criteria might be too stringent. Therefore we relaxed the filtering criteria, mainly with respect to quality and number of split reads, and then the RT could be detected carriers with 15 fold coverage for 5 out of the 7 (Table 4).
With 20 fold coverage all the RTs were detected by DELLY, but still many were of low quality and hence filtered out using the strict filtering criteria (Table 4). Relaxing the filtering criteria resulted in the detection of 6 of the 7 RT carriers (Table 4), which is just as good as with 30 fold coverage (Table 2). These results show that 20 fold coverage is a minimum coverage to detect RTs, and that filtering criteria should be adjusted to sequencing coverage.