Table 3.
Set | Gene count | LGD mutations in quads | Transmission pattern | Delta (Aut only − Sib only) | Delta P value | |||
None | Both | Aut only | Sib only | |||||
All | 18,455 | 19,602 | 4,671 | 5,197 | 4,921 | 4,813 | 108 | 0.1398 |
Genes with two UR LGD | 2,624 | 3,602 | 848 | 954 | 900 | 900 | 0 | 0.4981 |
Genes with one UR LGD | 4,538 | 3,114 | 757 | 840 | 809 | 708 | 101 | 0.0101 |
Split by IQ | ||||||||
Lower IQ | 1,586 | 1,586 | 383 | 437 | 418 | 348 | 70 | 0.0071 |
Higher IQ | 1,528 | 1,528 | 374 | 403 | 391 | 360 | 31 | 0.1489 |
Split by parent | ||||||||
Mother | 1,590 | 1,590 | 395 | 415 | 428 | 352 | 76 | 0.0031 |
Father | 1,524 | 1,524 | 362 | 425 | 381 | 356 | 25 | 0.1890 |
Split by length | ||||||||
Long | 1,557 | 1,557 | 384 | 412 | 427 | 334 | 93 | 0.0005 |
Short | 1,557 | 1,557 | 373 | 428 | 382 | 374 | 8 | 0.4034 |
Transmission in genes with one UR LGD mutation by functional category | ||||||||
FMRP | 128 | 128 | 32 | 33 | 38 | 25 | 13 | 0.0654 |
Chromatin | 62 | 62 | 15 | 18 | 18 | 11 | 7 | 0.1299 |
Embryonic | 298 | 298 | 59 | 82 | 94 | 63 | 31 | 0.0073 |
PSD | 245 | 245 | 57 | 64 | 69 | 55 | 14 | 0.1195 |
Essential | 258 | 258 | 69 | 63 | 75 | 51 | 24 | 0.0248 |
The first row shows transmission patterns for all of the 19,602 UR LGD mutations identified in parents of the 1,866 families for which both affected and unaffected siblings (quads) have been whole-exome sequenced (covering 18,455 genes). The four columns under the heading “Transmission pattern” give the numbers of UR LGD mutations transmitted to “None” of the children, to “Both” children, only to the affected child (“Aut only”), or only to the unaffected child (“Sib only”). We use the difference (delta) between the number of UR LGD mutations transmitted only to affected children and the number of UR LGD mutations transmitted only to unaffected children as a measure of overtransmission to the affected child. We test the significance of the delta against an empirically derived distribution through 10,000 iterations, randomly swapping the affected status of the two children within each family. Although there is a delta of 108 for all LGD mutations, it is not statistically significant (P = 0.1398). We then analyzed the subset of UR LGD mutations that occur in the 4,538 genes with exactly one UR LGD mutation in the SSC parents: 3,114 of these UR LGD mutations are in quads, and the delta in this smaller set of UR LGD mutations is 101; 809 are transmitted only to affected children, whereas 708 are transmitted only to the unaffected child. This delta is almost as large as the delta from all UR LGD mutations, and is significant (P = 0.0101). In contrast, the delta is 0 when we consider the UR LGD mutations in genes with exactly two UR LGD mutations. We then split the UR LGD mutations into roughly two equal halves independently, based on nonverbal IQ of the affected child, on the parent who carried the variant, and on the length of the gene measured as the load of UR synonymous variants. The overtransmission in each half is presented under the “split by IQ,” “split by parent,” and “split by length” subsections of the table. The most extreme difference is observed between long genes (delta = 93, P = 0.0005) and short genes (delta = 8, P = 0.4034), with the majority of overtransmission observed in the long genes. Most of the overtransmission is found in variants carried by the mother (delta = 76) relative to variants carried by the father (delta = 25), and in families with a lower-IQ affected child (delta = 70) relative to those families with a higher-IQ child (delta = 31). The lower section of the table shows the overtransmission of the UR LGD mutations in genes with one UR LGD mutation that are members of five functional classes: FMRP-associated, chromatin modifiers, embryonic, PSD (post synaptic density), and essential (10).