1
Supplement
Deletions of recessive disease genes: CNV contribution to carrier states and disease-causing alleles
Philip M. Boone,1 Ian M. Campbell,1Brett C. Baggett,1Zachry T. Soens,1Mitchell M. Rao,1 Patricia M. Hixson,1,2Ankita Patel,1,2Weimin Bi,1,2Sau Wai Cheung,1,2Seema R. Lalani,1,2,3,4 Arthur L. Beaudet,1,2,3,4,5Pawel Stankiewicz,1,2ChadA. Shaw,1,2James R. Lupski,1,2,3,4,*
1) Dept. of Molecular and Human Genetics, Baylor College of Medicine, Houston, TX, USA; 2) Medical Genetics Laboratories, Baylor College of Medicine; 3) Dept. of Pediatrics, Baylor College of Medicine; 4) Texas Children’s Hospital, Houston, TX, USA; 5) Dept. of Molecular and Cellular Biology, Baylor College of Medicine
I.Supplemental Methods
Assigning inheritance patterns to genes
To identify all genes associated with Mendelian disease and their respective inheritance patterns, we began with a list of all OMIM loci (13,304 loci listed in genemap.txt, downloaded from Loci were removed that were: 1) not genes (e.g. contiguous gene syndrome regions and mapped phenotypes with unknown molecular bases [denoted by “%” in OMIM]); 2) not associated with a phenotype; 3) only known to be mutated somatically;and 4)solely associated with phenotypes with the following characteristics:
- Unsubstantiatedgene-phenotype association(denoted by “?” in OMIM)
- Likely non-disease phenotypes (i.e.traits; denoted by “[ ]” in OMIM)
- Likely a low-penetrance/susceptibility association (denoted by“{}” in OMIM)
- Only described to be caused by translocation(s)
OMIM's application programming interface (API; was then used to obtain the inheritance pattern(s) for every available phenotype associated with each disease gene. If both recessive anddominant phenotypes, or a phenotype displaying both inheritance patterns, were associated with the gene, the gene was assigned as “rec/dom.” Werevised the resultant gene/inheritance list with input from other such published gene lists(Bell et al. 2011; Boone et al. 2013) andfrom personal communications with Dr. Stephen Kingsmore (Children’s Mercy Hospital, Kansas City, MO, USA; similar to the list of recessive disease genes published in (Saunders et al. 2012)) and Dr. Jonathan Berg (University of North Carolina, Chapel Hill, NC, USA; similar to the list of genes published in (Berg et al. 2013)),and made some assignments by manual review.
Note that our investigation required associating genes with inheritance patterns using phenotypes as an intermediary. Not all disease genes are associated with a single inheritance pattern. For example,among genes categorized as rec/dom, some are associated with two separate disorders displaying different inheritance patterns, while others are more consistent with semi-dominance (e.g.LDLR and familial hypercholesterolemia, MIM #143890). Furthermore, some recessive disease genes may be incorrectly described as being associated with dominant disease secondary to pseudodominance (Miksch et al. 2005), limited phenotyping (Plomp et al. 2004), or a failure of investigators to find two mutations at the same locus in an affected individual. Other more complex examples exist, for example the initial report of autosomal dominant congenital bilateral aplasia of the vas deferens (CBAVD; MIM #277180); upon further investigation, many patients are found to be compound heterozygotes consisting of CFTR coding mutations and the “5T” noncoding hypomorphic allele (Chillón et al. 1995). Nonetheless, classification by Mendelian inheritance pattern is still vital for most phenotypes and genes in a medical genetic setting; thus, Supplemental Table S1, along with other such published lists, may serve as a resource for medical genomics.
Statistical analysis of multiple carrier CNVs occurring in a single individual
To determine whetherthe observed numberof subjects harboring multiple carrier CNVs is consistent with independent assortment, we estimated the probability of any two or three genes being simultaneously deleted in the same individual by chance. This was done by calculating the probability of each of the possible pairs and trios of genes deleted at least once as part of aV8 Tier 1 heterozygous CNV containing a single recessive disease gene. The empirical V8 gene-specific frequency of deletion in our cohort was incorporated into our model, which summedthe product of these frequencies across all possible pairs and trios. Because the probabilities of quartets or higher-order combinations are very low compared to those of pairs and trios, and calculating the probability of quartets would be computationally intensive, we did not consider probabilities of combinations of four or more gene deletions in our model. We limited our analysis to CNVs containing exactly one recessive disease gene in order to reduce the effect of physical linkage between recessive disease genes. To determine the expected number of individuals with multiple recessive carrier CNVs, we parameterized a binomial distribution with a probability equal to our empirically determined probability of having any pair or trio of recessive disease genes deleted and a sample size equal to the size of our cohort. We also assessed the correlation between all pairs of recessive disease genes occurring in the same individual. Statistical analyses were completed using the R Statistical Computing Package (R Core Development Team).
Calculating Alu density per gene
We obtained coordinates of all annotated coding exons from RefSeq along with all annotated Alu elements using RepeatMasker. To overcome the problem of many genes having multiple isoforms, we selected the longest isoform by transcription start and stop and then by number of exons. For our analysis, we also included flanking intergenic regions of 47,603 bp (corresponding to 2 standard deviations above mean intron size) or until transcribed sequence of the next neighboring gene, whichever was smaller. Finally, we calculated the fraction of these intronic or flanking regions that consisted of Alu elements for each annotated RefSeq gene.
Identifying “consecutive” recessive disease genes
The following limitations to the list of 294 chromosomal regions containing “consecutive” recessive disease genes (Fig. 4; Supplemental Table S4) should be noted: The list of predicted potential recessive contiguous gene syndrome regions may grow as additional recessive disease genes are discovered and may shrink as dominant disease genes are identified. Our analysis did not exclude known, dominant genomic disorder regions in which no single causative disease gene has been identified, nor did it treat rec/dom disease genes in a sophisticated manner. Additionally, not all recessive conditions can be caused by deletion or duplication of their associated disease genes. Enhancers and other distal effects were not considered.
II. Supplemental Figures
Figure S1. Filtering strategy to identify potential carrier and disease-causing CNVs for all known recessive diseases. CNVs were filtered based on the following criteria: 1) Deletions; 2) Average probe ratio log2 ≤ -.415 for autosomal CNVs and X-linked CNVs in females (this is the expected value for a 50% mosaic loss; this threshold helps to eliminate both mosaic CNVs and poor quality calls) and ≤ -1 for X- and Y-linked CNVs in males (for the same reason as above); 3) CNVs that did not overlap an exon or flanking sequence were removed; 4) Only CNVs affecting at least one Mendelian disease gene listed in Table S1; 5) CNVs were excluded that overlapped any gene with dominant inheritance in Table S1. This was done to eliminate CNVs likely to cause dominant disease currently in a subject and thus be unrepresentative of carrier CNVs in the general, “healthy” population. CNVs including any gene associated with a disease phenotype with complex inheritance or other (e.g. digenic, parent-of-origin specific, etc.) inheritance patterns were also removed, owing to potential difficulty of interpretation; 6) The remaining CNVs were parsed into three tiers based on the likelihood that they are associated with recessive disease and not dominant disease: Tier 1) CNVs affecting only genes solely associated with recessive disease and non-disease genes; Tier 2) At least one gene in the CNV is solely associated with recessive disease, while one or more genes is associated with recessive and dominant disease (“rec/dom” genes). These CNVs, if heterozygous, are both potential carrier variants and may also be disease-causing variants; Tier 3) The CNV affects rec/dom disease genes and non-disease genes only. Thus, it is unclear whether a heterozygote for this variant would be a carrier, affected, or both; 7) All X- and Y-linked CNVs in males were assigned as hemizygous. Remaining CNVs with log2 > -2 (the theoretical log2 of a 1.5 copy loss; see Figs. S4 and S7) were called as heterozygous, and those with log2 ≤ -2 were assigned as homozygous deletions; 8) Heterozygous CNVs encompassing nine or fewer probes were subjected to additional QC to exclude calls unlikely to represent real deletions; this consisted of computational exclusion of deletions with high probe variance and visual curation of log2 plots for the remaining CNVs. Log2 plots of all homozygous and hemizygous CNVs were evaluated visually. Finally, all CNVs were reviewed to reconcile instances of overlapping CNVs in the same individual. One CNV was removed, and one reassigned zygosity, based on suspected gender mismatch with controls.
Figure S2. Percentage of subjects with a CNV deleting a recessive disease gene, grouped by zygosity, array version, and tier. See text for an explanation of the tier system.The proportion of subjects with a heterozygous deletion of any tier was significantly greater using the higher resolution and exon-targeted V8 CGH array (2,931 of 16,542 subjects; 17.7%) than the V7 non-targeted array (278 of 4,928 subjects; 5.6%) (p<.0001, two-tailed Pearson’s χ2 test).
Figure S3. Comparison of unique recessive disease genes deleted in the present study with those mutated in Bell et al.(2011).Bell et al.(2011) screened 437 recessive disease genes in a population of 104 individuals, most of whom were known carriers of or affected by recessive disease. 359 of these genes are “recessive” genes in Table S1 (i.e. not rec/dom, etc.). The overlap of recessive disease genes affected by any mutation (SNPs, indels, or CNVs), and CNVs specifically, in Bell et al. and those deleted by Tier 1 heterozygous CNVs in our cohort is displayed.
Figure S4. Tier 1 heterozygous deletions plotted by average probe log2 value, the number of probes in the CNV, and validation by PCR and FISH. CNVs not attempted to be validated, or with equivocal PCR results, are plotted in grey.
Figure S5. Heterozygous Tier 1 deletions ≥10 probes. a. Size spectrum of deletions ≥10 probes.When CNVs ≤ 9 probes are eliminated, the median size of a Tier 1 heterozygous deletion increases considerably, as expected (see also Supplemental Table S9).b. The ascertainment of unique genes affectedis unsaturated even after all subjects with a Tier 1 heterozygous deletion (x-axis) are tested, up to a total of 301 unique genes for CNVs ≥10 probes. c-e. These 301 recessive disease genes, when compared to the remaining known recessive disease genes, are significantly (c) farther from the nearest dominant gene (median 1.91 Mb vs. 665 kb; p<2.2x10-16), (d) larger (median genomic size 45.4 kb vs. 26.3 kb, p=2.6x10-10), and (e) have lower intronic and flanking Alu density (median 11.1% vs. 17.3%, p=7.2x10-10) (all statistics derived from the Wilcoxon rank sum test with continuity correction).
Figure S6. SNP array data for subject 2930. At least 665 Mb of AOH exist in this patient’s genome, consistent with consanguinity. LEPREL1 was located within one region of AOH (yellow shading), both explaining the origin of this subject’s homozygous deletion and providing additional evidence that our algorithm correctly called the zygosity of this CNV.
Figure S7. The distribution of unfiltered CNV log2 ratios suggests that a ≤ -2 log2 threshold for homozygous CNVs is appropriate. Data include 159,735 autosomal and X-linked (in females) CNVs prior to filtering.a. Full data. b. Zoomed-in view. The log2 cutoff implemented to identify autosomal (and X-linked in females) homozygous CNVs was ≤ -2 (dotted line). While the rationale for this value is discussed in the Figure S1 legend, we sought to determine empirically if this was a reasonable threshold. We hypothesized that the distribution of autosomal (and X-linked in females) deletion log2 values might be bimodal on account of overlapping distributions of homozygous and heterozygous deletions, and that a log2 cutoff at the trough between these peaks might maximize sensitivity and specificity when assigning ploidy to deletion CNVs. Indeed, deletions among our cohort follow a bimodal log2 distribution with a trough between the two local maxima very near -2. See also Figure S4.
III.Supplemental Tables
Table S1. All OMIM genes associated with Mendelian disease phenotypes, listed by inheritance pattern. Included as a separate file (.xls).
Table S2. CNVs passing all filtering steps. Included as a separate file (.xlsx).
Table S3. Deletion frequency by gene among V8 subjects. Included as a separate file (.xlsx).
Table S4. Genomic regions containing two or more recessive disease genes and no dominant or “rec/dom” disease genes.Included as a separate file (.xlsx).
1
Table S5. The recessive disease genes most commonly deleted by a Tier 1 heterozygous CNV.
Rank / Gene / Associated recessive condition(s) (OMIM ID) / No. times deleted, V8 / % deleted, V8 / 1 in __ / Exon(s) deleted / Likely pathogenic allele? / This/these CNV(s) previously reported?a / Published disease incidence1 / OCRLb / Lowe syndrome (#309000);
Dent disease 2 (#300555) / 314 / 1.90c / 53c / 314 subj: Ex 16/24 (in-frame) / Unlikely (too common in our cohort) / No / Lowe syndrome: “a few…per 100,000 births” (Lewis et al. 1993);
Dent disease: Approx. 250 families described(Devuyst and Thakker 2010)
2 / NKX2-6 / Persistent truncus arteriosus (#217095) / 262 / 1.58 / 63 / 214 subj: Ex 1/2;
48 subj: Ex 1-2/2
(noncoding by RefSeq gene model; coding by UCSC gene model) / ? / No / Truncus arteriosus: 3-5.6 per 100,000(Cifarelli and Ballerini 2005);
Single family with NKX2-6 mutations (Heathcote et al. 2005)
3 / LEPREL1 / High myopia with cataract and vitreoretinal degeneration (#614292) / 238 / 1.44 / 70 / 238 subj: Alternative ex 1/15 (noncoding) / Unlikely (too common in our cohort) / Yes (DGV) / Single family with LEPREL1 mutations (Mordechai et al. 2011)
4 / SLC12A3 / Gitelman syndrome (#263800) / 140 / 0.85 / 118 / 137 subj: Ex 9/26 (out of frame);
3 subj: Other exons / ? / Similar, but not exact (DGV) / 1/40,000 (Caucasians) (
5 / MKKS / McKusick-Kaufman syndrome (#236700); Bardet-Biedl syndrome 6 (#209900) / 95 / 0.57 / 174 / 91 subj: Ex 5/6 (in-frame);
4 subj: Other exons / ? / No / -
6 / KDM5Cb / Mental retardation, X-linked, syndromic, Claes-Jensen type (#300534) / 89 / 0.54c / 186c / 63 subj: Ex 14-15/24 (in-frame);
25 subj: Ex 13-15 (out of frame);
1 subj: Ex 15 (out of frame) / ? / No / Approx. 3% of X-linked intellectual disability (Jensen et al. 2005)
6 / STIL / Microcephaly, primary autosomal recessive, 7 (#612703) / 89 / 0.54 / 186 / 87 subj: Ex 3/27 (in-frame);
2 subj: Other exons / ? / No / Rare
8 / NRXN1d / Pitt-Hopkins-like syndrome 2 (#614325) / 86 / 0.52 / 192 / Various / Yes (one or more) / Yes (several in DGV, OMIM) / Rare
9 / NPHP1 / Nephronophthisis 1, juvenile (#256100); Senior-Løken syndrome 1 (#266900); Joubert syndrome 4 (#609583) / 82 / 0.50 / 202 / 80 subj: Whole gene;
2 subj: Other exons / Yes / Yes / -
10 / COL4A4 / Alport syndrome, autosomal recessive (#203780) / 78 / 0.47 / 212 / 67 subj: Ex 34/47 (in-frame);
11 subj: Other exons / ? / Similar, but not exact (DGV) / Approx 1/300,000 (Kang et al. 2010)
11 / OCA2 / Oculocutaneous albinism II and Albinism, brown oculocutaneous (#203200) / 52 / 0.31 / 318 / 27 subj:Whole gene;
25 subj: Other exons / Yes (one or more) / One intragenic deletion (OMIM) / 1/1,500-3,900 (Southern Africa), 1/38,000-40,000 (most populations) (Lewis 1993)
a DGV, HGMD (public access), and OMIM were queried.
b X-linked recessive.
c Frequency among all subjects analyzed by V8 array. Frequency among females is higher.
d Despite multiple reports of heterozygous mutations/deletions in this gene in individuals with neurological disorders, it has not been established in OMIM as a highly-penetrant dominant disease gene, and is thus was treated as a recessive disease gene in our analyses.
Rare, fewer than 100 cases described; Subj, subjects; ?, unclear; -, unknown.
1
Table S6. Tier 1 homozygous deletions.
Subject ID / CNV ID / Chr / Genes / Exons affected / Associated recessive phenotype(s) (OMIM ID) / Known mutation? / V8 Carrier freq / In DGV?163 / 3213 / 2p21 / SLC3A1*, PREPL / Whole gene / Hypotonia-cystinuria syndrome (#606407) / May be “I” (Régal et al. 2012) / 4/16,542 (any SLC3A1/PREPL del) / No
133 / 3214a / 2q13 / NPHP1*, LINC00116 / Whole gene / Nephronophthisis 1, juvenile (#256100); Senior-Løken syndrome 1 (#266900); Joubert syndrome 4 (#609583) / Yes (one or both alleles)(Konrad et al. 1996) / 82/16,542 / Yes
2930 / 3215 / 3q28 / LEPREL1* / Alternative exon 1 (noncoding) / High myopia with cataract and vitreoretinal degeneration (#614292) / No / 238/16,542 / Yes
2580 / 3216 / “ “ / “ “ / “ “ / “ “ / “ “ / “ “
867 / 3217 / “ “ / “ “ / “ “ / “ “ / “ “ / “ “
1236 / 3218 / “ “ / “ “ / “ “ / “ “ / “ “ / “ “
253 / 3219 / “ “ / “ “ / “ “ / “ “ / “ “ / “ “
1625 / 3220 / “ “ / “ “ / “ “ / “ “ / “ “ / “ “
* Recessive disease gene in Table S1.
a Compound heterozygous deletion (see Fig. 8b).
Chr = chromosome; min, minimum.
1
Table S7. Summary of hemizygous Tier 1 deletions.
XLR gene deleted / Associated recessive phenotype(s) (OMIM ID) / Hemizygous deletions (total) / Hemizygous deletions (V8) / Female carriers (V8 Tier 1 heterozygotes) / Hemizygous 1 in __ (V8)ARHGEF6 / X-linked mental retardation 46 (#300436) / 1 / 1 / 0 / 16542
ARSE / Chondrodysplasia punctata, X-linked recessive (#302950) / 1 / 1 / 2 / 16542
DMD / Duchenne muscular dystrophy (#310200); Becker muscular dystrophy (#300376); Dilated cardiomyopathy 3B (#302045) / 17 / 16 / 19 / 1034
HPRT1 / Lesch-Nyhan syndrome (#300322); HPRT-related gout (#300323) / 1 / 1 / 2 / 16542
IL1RAPL1 / X-linked mental retardation 21/34 (#300143) / 1 / 1 / 2 / 16542
L1CAM / Hydrocephalus due to aqueductal stenosis (#307000); MASA syndrome (#303350); CRASH syndrome (#303350); Hydrocephalus with Hirschsprung disease (#307000); Hydrocephalus with congenital idiopathic intestinal pseudoobstruction (#307000); Partial agenesis of corpus callosum (#304100) / 1 / 1 / 0 / 16542
MID1 / Opitz GBBB syndrome type I (#300000) / 1 / 1 / 1 / 16542
NDP / Norrie disease (#310600); X-linked exudative vitreoretinopathy (#305390) / 1 / 0 / 0 / -
OCRL / Lowe syndrome (#309000); Dent disease 2 (#300555) / 1 / 1 / 314 / 16542
OPN1LW / Colorblindness, protan (#303900); Blue cone monochromacy (#303700) / 16 / 8 / 0 / 2068
OTC / Ornithine transcarbamylase deficiency (#311250) / 2 / 1 / 2 / 16542
PAK3 / X-linked mental retardation 30/47 (#300558) / 2 / 2 / 8 / 8271
STS / X-linked ichthyosis (#308100) / 20 / 10 / 11 / 1654
UBE2A / X-linked mental retardation, syndromic, Nascimento type (#300860) / 2 / 2 / 1 / 8271
1
Table S8.Probe log2 statistics for heterozygous Tier 1 deletions.
Number of probes in CNV / 3 / 4 / 5 / 6 / 7 / 8 / 9 / ≤ 9 / ≥ 10Mean log2 / -.720 / -.703 / -.600 / -.583 / -.581 / -.670 / -.765 / -.669 / -.760
Median log2 / -.683 / -.684 / -.561 / -.532 / -.488 / -.644 / -.789 / -.626 / -.804
Number of CNVs / 838 / 521 / 448 / 301 / 114 / 52 / 62 / 2336 / 876
Table S9. Tier 1 deletions ≥10 probes.
Number / 876Mean sizea / 1,308 kb
Median size / 161 kb
Unique recessive disease genes in del / 301
Del described previously in DGVb / 506 (58%)
Dels with ≥ 2 recessive disease genes / 157 (18%)
a Size and all other statistics are based on the minimum deleted interval as assessed by aCGH.
b As assessed by 50% mutual overlap, irrespective of ploidy. DGV data obtained from
Del, deletion; DGV, Database of Genomic Variants
Table S10. Comparison of deletion carrier frequency in our cohortvs. point mutation carrier
frequency reported by Lazarin, et al.(2013).
Genea / Condition / Carrier frequency (%) / Carrier frequency (1 in __)Point mutations(Lazarin et al. 2013) / Deletion CNVsb / Point mutations(Lazarin et al. 2013) / Deletion CNVsb
DHCR7 / Smith-Lemli-Opitz syndrome / 1.47 / 0.060 / 68 / 1654
CNGB3 / Achromatopsia / 1.03 / 0.0060 / 98 / 16542
DPYD / Hereditary thymine-uraciluria / 0.93 / 0.13 / 108 / 752
ACADM / Medium chain acyl-CoA dehydrogenase deficiency / 0.88 / 0.030 / 113 / 3308
GAA / Pompe disease / 0.76 / 0 / 132 / -
PAH / Phenylalanine hydroxylase deficiency / 0.71 / 0.036 / 142 / 2757
HEXA / Hexosaminidase A deficiency / 0.64 / 0 / 156 / -
HEXA / Tay-Sachs disease / 0.56 / 0 / 177 / -
ACADS / Short chain acyl-CoA dehydrogenase deficiency / 0.44 / 0 / 229 / -
BBS10 / Bardet-Biedl syndrome, BBS1-related / 0.36 / 0.0060 / 281 / 16542
ATP7B / Wilson disease / 0.35 / 0.012 / 283 / 8271
GALT / Galactosemia / 0.35 / 0 / 288 / -
ASPA / Canavan disease / 0.33 / 0.042 / 304 / 2363
MUTYH / MYH-associated polyposis / 0.30 / 0 / 337 / -
BTD / Biotinidase deficiency / 0.24 / 0.0060 / 416 / 16542
BCKDHB / Maple syrup urine disease type 1B / 0.23 / 0.0060 / 434 / 16542
CLN3 / CLN3-related neuronal ceroid lipofuscinosis / 0.22 / 0.036 / 455 / 2757
FAH / Tyrosinemia type 1 / 0.22 / 0 / 465 / -
HADHA / Long chain 3-hydroxyacyl-CoA dehydrogenase deficiency / 0.20 / 0 / 510 / -
ARSA / Metachromatic leukodystrophy / 0.19 / 0.048 / 527 / 2068
SMPD1 / Niemann-Pick disease, SMPD1-associated / 0.18 / 0.0060 / 555 / 16542
ACADVL / Very long chain acyl-CoA deficiency / 0.17 / 0 / 576 / -
MCOLN1 / Mucolipidosis IV / 0.17 / 0.012 / 597 / 8271
CTNS / Cystinosis / 0.16 / 0.18* / 608 / 551*
BLM / Bloom syndrome / 0.16 / 0 / 613 / -
SMPD1 / Niemann Pick disease type A / 0.16 / 0.0060 / 617 / 16542
TPP1 / TPP1-related neuronal ceroid lipofuscinosis / 0.16 / 0 / 632 / -
IVD / Isovaleric acidemia / 0.12 / 0.0060 / 832 / 16542
BBS10 / Bardet-Biedl syndrome, BBS10-related / 0.10 / 0.0060 / 988 / 16542
NEB / NEB-related nemaline myopathy / 0.091 / 0.024 / 1091 / 4136
LAMB3 / Herlitz junctional epidermolysis bullosa, LAMB3-related / 0.089 / 0 / 1123 / -
CBS / Homocystinuria caused by cystathionine beta-synthase deficiency / 0.083 / 0.0060 / 1210 / 16542
ASS1 / Citrullinemia type 1 / 0.079 / 0 / 1266 / -
PPT1 / PPT1-related neuronal ceroid lipofuscinosis / 0.076 / 0 / 1318 / -
HSD17B4 / D-bifunctional protein deficiency / 0.063 / 0.0060 / 1583 / 16542
AIRE / Polyglandular autoimmune syndrome type 1 / 0.057 / 0.0060 / 1757 / 16542
NBN / Nijmegen breakage syndrome / 0.051 / 0.012 / 1976 / 8271
POMGNT1 / Muscle-eye-brain disease / 0.045 / 0 / 2247 / -
GCDH / Glutaric acidemia type 1 / 0.045 / 0 / 2247 / -
PROP1 / PROP1-related combined pituitary hormone deficiency / 0.039 / 0 / 2549 / -
SGCB / Limb-girdle muscular dystrophy type 2E / 0.038 / 0 / 2621 / -
SGCA / Limb-girdle muscular dystrophy type 2D / 0.032 / 0.0060 / 3166 / 16542
TH / Segawa syndrome / 0.025 / 0 / 3932 / -
CPT1A / Carnitine palmitoyltransferase 1a deficiency / 0.013 / 0 / 7864 / -
LAMC2 / Herlitz junctional epidermolysis bullosa, LAMC2-related / 0.0064 / 0.048* / 15727 / 2068*
SACS / ARSACS / 0 / 0.13* / - / 788*
CLN5 / CLN5-related neuronal ceroid lipofuscinosis / 0 / 0 / - / -
CLN8 / Northern epilepsy / 0 / 0.030* / - / 3308*
ALDH3A2 / Sjögren-Larsson syndrome / 0 / 0.024* / - / 4136*
Average / 0.26 / 0.019
aGenes are those 1) tested by Lazarin et al. (2013), 2) listed as recessive (i.e. not rec/dom, etc.) in Supplemental Table S1, and 3) having exonic probe coverage on the V8 CGH array. Lazarin, et al. reported the phenotypes, but not genes, tested in most cases; the genes tested in these cases are assumed from the phenotype. All genes are autosomal.