Educating children with 22q13 deletion syndrome

David is tolerant of my picture taking

If you ask me what is wrong with David, I don’t think I can answer the question without making a laundry list of problems, deficits and other issues. He has 22q13 deletion syndrome – that is what’s wrong.  Some of his problems almost killed him.  Someday, that risk may return.  In the meanwhile, his biggest problem is learning new knowledge and skills.  Can we make learning more efficient for our children?  To investigate this question, we need to understand more about how individual and groups of genes can affect the brain.

I have written blog posts on specific issues (cancerulcerative colitishypotonia) and other issues in the context of individual genes. Individual genes can impact many parts of the body.  A single gene may have multiple functions, an effect called “pleiotropy” (plahy-o-truh-pee).  We are often concerned with the primary impact of a gene more so than the secondary issues.  For example, CTFR is a cystic fibrosis gene. Its impact on the lungs is very damaging, but it also impacts other parts of the body.

Many 22q13 deletion syndrome genes are pleiotropic.  We must look carefully at all the potential effects of a gene.  Why? Because, so many genes are lost in 22q13 deletion syndrome that subtle effects can add up. When I read the scientific papers on a gene, I spend a lot of time comparing the subtle effects of this gene with all the others.  I look for cases where multiple genes have subtle effects on the same organ. By definition, 22q13 deletion syndrome is a chromosomal deletion syndrome.  It is not a monogenic syndrome as some have suggested.  I recommend using the name “Phelan-McDermid syndrome” if you want to combine SHANK3 mutation syndrome with 22q13 deletion syndrome. See: Introduction to 22q13 deletion syndrome and How to fix SHANK3.

Pleiotropic effects come in two flavors.  Either the gene has one function, but in different parts of the body, or the gene can do more than one function.  CTFR, is not a 22q13 deletion syndrome gene, but it provides a useful example.  It is the most common cystic fibrosis gene.  CTFR is involved in making secretions (fluids used in the body). CTFR mutations are most important in the lungs, but the gene also causes faulty secretions in the digestive tract, and elsewhere.

CELSR1 is a 22q13 deletion syndrome gene.  Over 40% of our children are missing this gene.  Like CTFR, one function affects many different parts of the body.  If one copy of CELSR1 is mutated, the most serious result is a neural tube closure defect (e.g., spina bifida or other spinal cord problems).  Mouse studies of Celsr1 show that it participates in helping cells organize into physical patterns so that cells can operate as a group. Celsr1 is involved in early development by organizing certain cells into functioning tissues (Feng et al, 2012).  However, the central role of CELSR1 in adult brain function was only discovered last year (Schafer et al, 2015).

During development, Celsr1 mutations can interfere with the organization of many different cells (Boutin et al, 2014). For example, ventricles are nourishing fluid lakes inside the brain. Cilia, cells with tiny hairs, line the ventricles of the brain and stir the cerebrospinal fluid (CSF) along its path through the ventricles. Stagnation of the CSF fluid is dangerous. Disordered cilia from Celsr1 mutations cause inefficient motion. In humans, stagnant CSF may have accumulating impact over years.

Very different cells with cilia are used in the ear to hear sounds. CELSR1 mutations disrupt the orientation of “outer hair cells,” responsible for hearing at low sound levels. Cilia are responsible for other body functions, as well, like keeping the airways clear and digesting food. Given its impact on different organs, CELSR1 is pleiotropic.

When someone asks me what is wrong with David, one of the first things I say is he struggles to learn.  David is aware and interested in his environment, but he knows trying to learn anything new is difficult.  My last blog discussed SHANK3 and its impact on learning that involves the ventral striatum in the brain.  Mutations in CELSR1 disrupt a different kind of learning, learning that is unique to the hippocampus of the brain.  Only two areas of the brain are able to grow new neurons.  One of these areas feeds the new neurons into the dentate gyrus of the hippocampus and has a subtle, but critical effect on learning . Mutation of rodent Celsr1 disrupts the orientation of these new neurons in the hippocampus.  This disruption interferes with building proper connections (Schafer et al, 2015).  Thus, children with deletions greater than 6 Mbase are likely have problems with “pattern separation,” the very subtle learning process that prevents new learning from interfering with old knowledge (Johnston et al, 2015).  The concept of pattern separation has arisen through complex mathematical learning models.

How can we take this information on CELSR1 and translate it into treatments?  There are several paths.  First, we would like someone to study a mouse missing one complete copy of Celsr1 (heterozygous knockout mouse) to make sure the effects seen in the current mice are not simply mutation effects (see Gene deletion versus mutation).  There is a scientist in Belgium, Fadel Tissir, who has a “null” mouse (no copies of the gene), but has not reported any studies with a heterozygous knockout mouse yet.  Second, we need to find out about outer hair cell loss and its potential impact on hearing.  Perhaps someday we can arrange for a medical histologist to examine the cochlea (hearing organ) from a 22q13 deletion syndrome organ donor. Third, we need genetic testing (sequencing is best) on all patients missing CELSR1 to identify cases where the remaining CELSR1 gene is mutated.  A mutation on the remaining CELSR1 gene could unmask recessive traits in a way that may be very informative.

Finally, if we really want to understand our children’s learning problems, we need to: 1) engage people involved in computational models of learning, and 2) study patients with interstitial deletions.  Interstitial deletions will allow us to study genes in better isolation.  These patients may be higher functioning, which is ideal for careful testing (see How to fix SHANK3). As the new studies start to bear fruit, we can then use the results to target our teaching methods.  Right now parents, teachers and our children are frustrated with how poorly our children learn. Imagine how much better it will be when we know how to maximize learning for each child based on their genetic report.  I dream of seeing the next generation of children with 22q13 deletion syndrome having all the benefits we never had for David.  Let’s take lesson planning into the 21st century!


Previous blogs

How to fix SHANK3

Have you ever met a child like mine?

How do I know which genes are missing?

Mouse models
Science Leadership
How can the same deletion have such different consequences?
22q13 and the hope of precision medicine
22q13 Deletion Syndrome: hypotonia
Understanding gene size
Gene deletions versus mutations: sometimes missing a gene is better.
Is 22q13 deletion syndrome a ciliopathy?
Understanding translocations in 22q13 deletion syndrome: genetics and evolution
Understanding deletion size
Can 22q13 deletion syndrome cause ulcerative colitis?
Can 22q13 deletion syndrome cause cancer?
22q13 deletion syndrome – an introduction


Have you ever met a child like mine?

Sometimes, David likes to be held.
Sometimes, David likes to be held.

Jannine Cody, the parent/scientist who studies 18q deletions, says that since every deletion is different, every child with a deletion is different.  At the PMS family conferences we met other children with 22q13 deletion syndrome who, at the time, had striking similarities with David.  These children had chromosome 22 deletions of various sizes, and similar children did not always seem to have the same size deletions.  We know now that genes are not distributed equally along 22q13, so children with small deletions can be quite different from each other, and children with large deletions can be quite similar (see Understanding deletion size).  We also know there are good scientific reasons to expect differences (see How can the same deletion have such different consequences?).  Some things are pretty obvious after a while.  The kids who could not walk or talk generally had larger deletions.  Those with larger deletions also had many more medical problems.  Obviously, more genes lost means more problems.  Regular readers of this blog have seen evidence of why it is very important to know which genes are missing (see How do I know which genes are missing?).

Some people feel that research on 22q13 genes should be done one gene at a time, starting with SHANK3.  I am not a big proponent of this approach, since it ignores a lot of research already done on ARSA, MAPK8IP2, CHKB, CPT1B,  PANX2, ALG12, BRD1, SULT4A1 and other genes known to cause disorders in humans, mice or both.  The one gene-at-a-time approach also slows research by making one gene sound much more important than others.  It seems to me if we spend 5 to 10 years on each gene, we are doomed to spending 500 to 1,000 years. If that sounds pretty absurd, well, it is.  Maybe it will only take 200 years to do it this way.  That still seems too long to me.  That is why I recommend the scientific program be managed by someone with a deep understanding of science leadership (see 22q13 deletion syndrome and science leadership).  The “SHANK3 or bust” research program has succeed in some ways.  Recently, after about a dozen mouse models of Shank3, there is a new mouse with the first complete deletion of the gene. All the other mice were various examples of gene mutation.  As we know, the effects of mutation (or removing part of the gene) can be very different from deletion (see Gene deletion versus mutation: sometimes missing a gene is better). This is critically important! The main reason for supporting Shank3 mouse research is the argument that most (not all) patients are missing the SHANK3 gene entirely.  Thus, it is SHANK3 deletions that make the research important to our families. (Note that mouse Shank3 mutation research has a very separate goal: understanding how mutations might contribute to general forms of autism.)

So, we now have a real Shank3 deletion mouse and everyone is very excited about it (Mouse Model of Autism Offers Insights to Human Patients, Potential Drug Targets).  Of course, be skeptical of what the university PR team says (see Mouse models).  Let’s take a look at this first-ever complete Shank3 knockout mouse. First off, the major finding is that this mouse is different from the many mutation mouse models.  No one should be surprised.  What is surprising is that you have to completely wipe out 100% of Shank3 to see a measurable difference between these mice and normal mice. Even more shocking is that these mice are walking around, playing with other mice, eating, talking mice talk (ultrasonic sounds) with no shank3 whatsoever in their bodies!  The mice missing 100% of Shank3 are different from other mice, but mice missing 50% are not different in any measurable way. Note that humans with 22q13 deletion syndrome are missing only one of the two genes and best evidence is that they have lost only about 25% of their shank3 protein (See this research paper).

So, is there something wrong with the mouse study?  Are mice just way different from humans, or is there another explanation?  Maybe it all makes sense.  Have you ever met a human missing all of SHANK3 and only SHANK3?  The complete knockout Shank3 mouse is best compared with a person like that, someone who is not missing any other genes and has no known mutations.  It is not good enough to have someone with a “small deletion”, since there is strong evidence that adjacent genes impact brain function.  This mouse models SHANK3 deletion.  I have met only one person who seems to fit this description.

Phelan McDermid syndrome is characterized by developmental delays, moderate to severe intellectual disability, little or no expressive language, and infant hypotonia (floppy baby syndrome).  Some people argue that the syndrome is also characterized by a high incidence of autism spectrum disorder, although some top scientists disagree.  The person I met was probably never a floppy baby, has practically normal speech, and that person has no evidence of autism.  Rather, the person I met has some problems with coordination, has a great difficulty learning and is socially a wonderful person to meet and engage with, perhaps to a fault.  Tragically, like all of our children, that person will never navigate the world well enough to live an independent life.

In summary, when I read the scientific paper on the complete Shank3 knockout mouse, what struck me was how many tests the complete, 100% knockout mouse passed without demonstrable evidence of a problem. Mice missing one copy are normal in almost every test.  Mice missing both copies are not “normal”,  but clearly, even these mice are nothing like my son.

How important is SHANK3?  It is impossible to make that judgement based on only one clinical case.  The person I met has lost all independence for that person’s entire life. That is very important.  Moreover, it is tragic.  But for 95% of families, 22q13 deletion syndrome comes with the full set of core features of 22q13 deletion syndrome.  David cannot tell me when he feels sick, where it hurts, or if he was mistreated in his group home. It took him 6 years to overcome his floppy baby syndrome enough to walk and three more years before he could eat by mouth.  His autism-like features interfere with social contact.

As of now, the most parsimonious explanation of what we know is that SHANK3, alone, does not produce the core features of 22q13 deletion syndrome.  It is a contributor in most, but not all, cases.




Previous blogs

How do I know which genes are missing?
Mouse models
Science Leadership
How can the same deletion have such different consequences?
22q13 and the hope of precision medicine
22q13 Deletion Syndrome: hypotonia
Understanding gene size
Gene deletions versus mutations: sometimes missing a gene is better.
Is 22q13 deletion syndrome a ciliopathy?
Understanding translocations in 22q13 deletion syndrome: genetics and evolution
Understanding deletion size
Can 22q13 deletion syndrome cause ulcerative colitis?
Can 22q13 deletion syndrome cause cancer?
22q13 deletion syndrome – an introduction


How do I know which genes are missing?

David March 5David decided to stand still so I could
take his picture.  Thanks David!

I often get asked to interpret a genetic report.  I am not a clinician of any sort. I never even had a genetics course.  But, there are certain skills one must pick up along the way to study which genes are missing in 22q13 deletion syndrome and what they do.  So, I have some idea how to read a report, just enough to accomplish my goals.

If you have a genetic report, here are some ways you, too, can figure out which genes are missing.  David only had a FISH test 15 years ago, so I don’t know how many genes he is missing.  You may be in much better shape. If you had an array or sequence done on your child, you should have enough information to find out which genes are missing.  Here is how.

Expert mode

The experts use the technical part of a genetic report to see what has been deleted. The technical part will often look something like this:
hg17  arr 22q13.33 (48,230,183-49,523,149) X1
This example is a simple terminal deletion of chromosome 22 (not a real case).  It says that part of chromosome 22 only has one copy (X1).  The nomenclature can get complicated if there are multiple genetic problems.  The nomenclature is outlined here:

Once you know which base-pairs are missing (48,230,183-49,523,149) and which genome assembly was used to make the measurement (hg17), you can go to the UCSC Genome Browser and get gobs of information about the deletion.  The Genome Browser is complicated to use, but I use it often. 


Sometimes the report provides a complete list of the missing genes.  “Complete” is relative to certain things.  For instance, most gene array studies cannot detect the presence or loss of the last gene on the chromosome, RABL2B.  This gene is too similar to RABL2A and no one thinks RABL2B is important enough to make a special effort to detect it. (I happen to disagree. See my blog Is 22q13 deletion syndrome a ciliopathy.) Otherwise, the gene list is useful.  If the only listed genes are SHANK3 and ARSA, the gene list is worthless.  It means the geneticist either did not want to look up, or share with you, the other genes that are missing.

End gene or deletion size

If the deletion is a terminal deletion, you can use the deletion size to look up the genes on my list (below).  If the report says something like, “from PLXNB2 to ACR” or “distal to” PLXNB2, then you can find PLXNB2 on the list and it becomes the marker for where the deletion starts. Terminal deletions start from the “end” gene and go up the list (down in gene number on my list).

Go to the Foundation’s web site?

If you had 18q deletion syndrome, there is a site that was created for parents to look up the genes that are missing and what those genes do.  No such site exists for 22q13.

Ask your geneticist

I have seen some very nice genetics reports with a complete list of genes deleted.  This is rare, but you might try asking for such a document.

The list

Here is a list of the genes that covers the last 9,300 Kb (9.3 Mb) of 22q13.  That is about the size of the largest known terminal deletion.  The list comes with a few caveats.  First, deletions can be messy.  Sometimes lots of genes are gone, then part of the subsequent gene is gone.  People don’t count half a gene.  They usually add the partial gene to the gene count. Micro arrays (gene chips) sample the genes every so many kilobases (Kb).  It is not a continuous readout of the DNA.  Thus, in most cases, you don’t know the exact position of the break.   DNA includes many things in addition to genes. It encodes things called microRNAs, and the DNA has essential gene promoter, inhibitor and enhancer regions within genes and between genes.  Thus, a gene list tells only part of the story.  Finally, I found an error in my list this past week, so let me know, but please don’t complain if this has more errors!  I will correct this page as I find errors.  However, I emphasize that this list comes without any warranty whatsoever.  I take no responsibility for its use.  It is part of a blog to educate parents.  It is not a tool for legal, medical or any other practice. Ok?
[updated/corrected 6 March 2016]
[updated/corrected 12 March 2016]
[updated/corrected 17 June 2016]
[updated/corrected 5 December 2016]
[updated/corrected 7 April 2017]
Table now shows the smallest distance to the gene. For example, a terminal deletion of size 196.0 Kbase will delete all the genes before ARSA and damage part of ARSA.  Table is now limited to 9.3 Mbase (9,300,000) which is the largest recorded terminal deletion.

#       Gene       Deletion size (Kbase) 
1       RABL2B        34.81 
2       ACR           78.11 
3       SHANK3        85.17 
4       ARSA         195.14 
5       MAPK8IP2     206.92 
6       CHKB         235.47 
7       CPT1B        240.05 
8       SYCE3        255.56 
9       KLHDC7B      267.45 
10      ODF3B        286.65 
11      TYMP         288.47 
12      SCO2         292.86 
13      NCAPH2       295.00 
14      LMF2         310.78 
15      MIOX         328.43 
16      ADM2         332.03 
17      SBF1         343.44 
18      PPP6R2       373.38 
19      DENND6B      491.41 
20      PLXNB2       539.63 
21      MAPK11       548.08 
22      MAPK12       557.16 
23      HDAC10       567.21 
24      TUBGCP6      573.90 
25      SELO         600.85 
26      TRABD        618.90 
27      PANX2        638.18 
28      MOV10L1      656.85 
29      MLC1         733.11 
30      TTLL8        763.84 
31      IL17REL      809.78 
32      PIM3         854.39 
33      CRELD2       895.76 
34      ALG12        900.01 
35      ZBED4        928.39 
36      BRD1         995.26 
37      C22orf34     1,160.96 
38      FAM19A5      2,067.65 
39      TBC1D22A     3,642.77 
40      CERK         4,080.21 
41      GRAMD4       4,180.53 
42      CELSR1       4,281.30 
43      TRMU         4,484.77 
44      GTSE1        4,487.66 
45      TTC38        4,543.18 
46      PKDREJ       4,555.11 
47      CDPF1        4,570.18 
48      PPARA        4,620.06 
49      PRR34        4,764.32 
50      WNT7B        4,841.34 
51      ATXN10       4,973.16 
52      FBLN1        5,287.14 
53      RIBC2        5,385.97 
54      SMC1B        5,404.90 
55      FAM118A      5,508.30 
56      UPK3A        5,522.59 
57      KIAA0930     5,577.70 
58      NUP50        5,630.45 
59      PHF21B       5,809.48 
60      ARHGAP8      5,955.68 
61      PRR5         6,080.79 
62      LDOC1L       6,320.17 
63      KIAA1644     6,505.62
64      PARVG        6,610.00 
65      PARVB        6,649.55 
66      SAMM50       6,821.94 
67      PNPLA3       6,870.90 
68      PNPLA5       6,926.46 
69      SULT4A1      6,955.97 
70      EFCAB6       7,248.33 
71      MPPED1       7,310.62 
72      SCUBE1       7,475.08 
73      TTLL12       7,631.34 
74      TSPO         7,655.23 
75      MCAT         7,675.07 
76      BIK          7,688.76 
77      TTLL1        7,729.13 
78      PACSIN2      7,942.26 
79      ARFGAP3      7,961.07 
80      A4GALT       8,123.48 
81      ATP5L2       8,177.87 
82      CYB5R3       8,173.96 
83      RNU12        8,203.08 
84      POLDIP3      8,203.60 
85      SERHL2       8,260.06 
86      RRP7A        8,298.67 
87      SERHL        8,305.91 
88      NFAM1        8,386.07 
89      TCF20        8,603.03 
90      CYP2D6       8,688.56 
91      NDUFA6       8,727.69 
92      SMDT1        8,735.12 
93      FAM109B      8,739.03 
94      NAGA         8,747.64 
95      WBP2NL       8,785.70 
96      SEPT3        8,828.88 
97      CENPM        8,871.30 
98      TNFRSF13C    8,891.65 
99      SHISA8       8,903.80 
100     SREBF2       8,911.16 
101     CCDC134      8,992.58 
102     MEI1         9,019.01 
103     C22orf46     9,124.62 
104     NHP2L1(SNU13)9,129.56 
105     XRCC6        9,154.43 
106     DESI1        9,197.37 
107     PMM1         9,228.58 
108     CSDC2        9,241.80 
109     POLR3H       9,273.98 
110     ACO2         9,289.48 

Check out my earlier blogs (below) to learn how many of your friends are missing a similar number of genes and which genes might be important.  If you find this information valuable, please leave me a positive comment, or repost the blog elsewhere (e.g., Facebook).  Many thanks to each parent who has shared his/her child’s information and shared their own very personal stories. Your contributions and feedback helps me feel that I am not alone in the quest to make the world a better place for our kids.


Previous blogs

Mouse models
Science Leadership
How can the same deletion have such different consequences?
22q13 and the hope of precision medicine
22q13 Deletion Syndrome: hypotonia
Understanding gene size
Gene deletions versus mutations: sometimes missing a gene is better.
Is 22q13 deletion syndrome a ciliopathy?
Understanding translocations in 22q13 deletion syndrome: genetics and evolution
Understanding deletion size
Can 22q13 deletion syndrome cause ulcerative colitis?
Can 22q13 deletion syndrome cause cancer?
22q13 deletion syndrome – an introduction

Comment la même suppression peut-elle avoir une telle variété de conséquences ?

[English version]

Surviving the Neonatal ICU

L’auteur, Andy, et sa cousine ont la même chromosome hybride 22 qui provoque le syndrome de la suppression 22q13.

Leurs enfants ont la même suppression mais un des enfants d’Andy est mort et l’autre, David, a failli mourir. La cousine n’a pas eu cette expérience et sa fille, contrairement à David, n’a pas pris 6 ans à apprendre à marcher, 9 ans à manger oralement et elle peut parler avec les phrases courtes. C’est la question la plus posée par les parents d’enfants atteints.

Si SHANK3 n’est pas la cause de la suppression 22q13, pourquoi y a-t-il un enfant avec uniquement la mutation SHANK3 qui ne sait ni parler ni marcher ? Et il y en a d’autres qui savent très bien marcher et parler. Si vous acceptez la dilemme que des suppressions similaires peuvent provoquer des résultats très différents, il n’y a plus l’argument du gène favori (SHANK3 aujourd’hui, demain autre chose probablement.)

Pour expliquer ces différences il y a plusieurs réponses :

1. Perte de l’hétérozygotie : Des petites erreurs génétiques ont lieu tout le temps pendant le développement et dans la vie adulte. Facteurs environnementaux – radiation, infections, coups de soleil, toxines. Le fait d’avoir les paires de gènes de deux parents différents nous protège des erreurs génétiques graves. Lorsqu’il manque des gènes à cause d’un syndrome de suppression, on n’a qu’une copie d’un gène alors il y a l’occasion parfaite pour des erreurs non-corrigées. Si ce gène est endommagé par la suite, les conséquences peuvent être graves. L’erreur peut être globale (corps entiers et détectable avec les tests génétiques) ou local (limité à une partie du corps ou du cerveau.) Quand l’erreur est locale elle n’est pas détectable et devient une différence inexpliquée.

2. Impression (ou empreinte ?) : Quand un des gènes est éteint comme le syndrome d’Angelman. Si la seule copie qui reste du gène s’éteint ça va provoquer des problèmes qui ne seront pas expliquées.

3. Impact de la mutation du gène : Souvent un gène qui est supprimé a moins d’impact qu’un gène qui a subi une mutation. Comme dans le cas des cellules de cancer ; on préfère qu’elles meurent au lieu de pousser avec des gènes modifiés. Si une suppression chromosomique ne frappe qu’une partie du gène, le gène peut commencer à créer des protéines qui empêchent le fonctionnement normal de la cellule et la petite suppression provoque des grands problèmes.

4. Les combinaisons de gènes : Quand il y a plusieurs suppressions et mutations on cumule les erreurs génétiques et chacune fait sa contribution. On a parfois des gènes qui, seuls, n’ont pas d’impact lors de leur suppression/mutation mais avec certains autres peuvent avoir des conséquences beaucoup plus importantes. On ne comprend pas encore ces combinaisons alors on utilise des termes comme ‘antécédents génétiques’ en attendant. La différence génétique principale entre David et la fille de la cousine d’Andy vient de la femme d’Andy et du mari de sa cousine ; des différences dans leurs antécédents génétiques.

5. Mosaïcisme et mutations somatiques : Pendant le développement les erreurs génétiques peuvent avoir lieu dans une petite partie du cerveau. Ces erreurs peuvent expliquer les variations comme les difficultés d’apprentissage. Ces mutations inaperçues peuvent avoir des incidences plus lourdes lorsqu’elles interagissent avec les 30 à 100 gènes manquants. L’impact de SHANK3 peut être amplifié par ces gènes modifiés ou perdus dans des parties spécifiques du cerveau.

6. Les régulateurs génétiques: Le projet génétique ENCODE tente de trouver tous les morceaux régulateurs des gènes de l’ADN. La plupart de l’ADN est composée de régulateurs de gènes. C’est plus facile à comprendre quand vous vous rendez compte que les cellules de foie, de peau, des intestins, du cerveau, ont toutes les mêmes gènes. Ce qui change est les gènes qui sont actifs ou inactifs. Les cellules de peau savent qu’elles sont cellules de peau et n’utilisent que les gènes de cellules de peau. L’ADN est réglé dans chaque tissu à associer la signature génétique nécessaire à fabriquer le tissu. Les suppressions chromosomiques 22 suppriment également les régulateurs de gènes. Les régulateurs sont très difficiles à repérer et pourraient être la cause de ces différences inexpliquées.

Compte tenu de la complexité et de nombreuses possibilités de variation inexpliquée , nous pouvons commencer à apprécier que la connaissance de la taille de suppression d’un individu ne fournit pas toutes les réponses. Cependant, grâce à des outils modernes, il existe des moyens pour étudier les effets de la taille de la suppression même avec une telle variabilité. Ces outils peuvent être utilisés pour démêler les gènes qui contribuent à chaque problème médical. Cela exige un engagement sérieux des parents à pousser les chercheurs et le personnel médical vers ces recherches, tirant pleinement parti des rapports génétiques. Trop l’accent sur un gène préféré entrave le progrès scientifique et médical. Ceux qui travaillent sur un autre chromosome, syndrome de délétion (suppression 18q syndrome) ont étudié leur syndrome à bon escient au cours des 50 dernières années. Ils se tournent vers la médecine scientifique pour les symptomes de 18q (Voir Création d’ anomalies chromosomiques). Les personnes 18q ont développé une feuille de route, que les gens atteints du syndrome de délétion 22q13 peuvent facilement suivre (Voir Conséquences du chromosome 18q suppression). J ai travaillé durement dans mes dernières recherches pour faire ce point , mais rien est plus convaincant que de voir d’autres prendre les devants avec une telle clarté et tel engagement. Pourquoi n en avons-nous pas profité ? La seule explication que je peux trouver est que la communauté du syndrome de délétion 22q13 manque de personnes qualifiées, dirigeant scientifique impartiale. Il y a un problème assez évident, avec des conséquences très tristes. Il n’y a pas plus de traitements pour David aujourd’hui qu’il ya 30 ans. Nous savons quels gènes sont portés disparus et pour beaucoup d’entre eux, nous savons ce qu’ils font (Voir Comment savons-nous quels sont les gènes 22q13 suppression: l’espoir de la médecine de précision). Ce que nous ne semblons pas savoir est comment rendre le travail de la science meilleur pour le bien de nos familles.


[My thanks to Betty Sepré for doing this translation.  That said, I take responsibility for any errors in typing, translation or content. Feel free to contact me with corrections. –  arm22q13]

How can the same deletion have such different consequences?

David's deletion was the same as his cousin's, yet David's deletion has had more severe consequences.
David at 3 days of age: David’s deletion size is exactly the same as his cousin’s, yet the deletion had more severe consequences.

My cousin and I inherited the same bum chromosome.  Somewhere back in family history, tiny bits of chromosomes 19 and 22 got swapped (See Who is arm22q13?). I am the proud owner of a hybrid and unhelpful chromosome 22 that causes 22q13 deletion syndrome (See Understanding translocations in 22q13 deletion syndrome: genetics and evolution).  My equally unfortunate cousin has the same chromosome.  Perhaps her’s is slightly different from mine, but it is not likely very different considering how much we know about its common source.  In spite of having the exact same deletion, our children turned out very differently.  Both of my kids who received this chromosome were failure to thrive babies.  One died and the other one almost died (See photo of David).  My cousin had no such experience.  Her daughter has 22q13 deletion syndrome, but unlike David, she did not spend 6 years learning to walk, 9 years before oral feeding and she can talk in short sentences.  Why has virtually the exact same deletion had such different consequences?

This is the question I am asked the most by other parents of children with 22q13 deletion syndrome. Why does one child with a much larger deletion talk, while another child with a much smaller deletion seem much worse?  To be sure, this disparity in phenotype manifestation is very real. It is all around and we should not be swayed by phenotype-genotype deniers, if any exist.  The reality of large phenotype variability is important because it address another often asked question.  If SHANK3 is not what causes 22q13 deletion syndrome, why is there a child with only a SHANK3 mutation that can’t walk or talk? Of course, other children with only SHANK3 mutations walk and talk, some remarkably well.  Once you accept the dilemma that similar deletions can have very different outcomes, this argument for a favorite gene (SHANK3 today, probably some other gene in the future) goes away.

So, how do we explain these dramatic differences?  The answer is, there are many answers.  There are so many ways that similar deletions can have very different outcomes, that it takes a catalog of explanations to cover them.  Here we go.

  1. Loss of heterozygosity (a.k.a., hemizygosity).  Small genetic errors occur all the time during development and in adulthood.  Environmental factors from cosmic radiation to infections, sunburn to environmental toxins, can create small errors.  Cells have mechanisms to repair errors, but one important hedge against serious genetic errors is the fact that we carry two of every gene (one from mom and one from dad).  When a person or a tissue in the body has only one copy of a gene, there is a unique opportunity for uncorrected errors.  22q13 deletion syndrome is the loss of some or many genes on one chromosome.  This creates hemizygosity (“half as many copies”) for those genes.  Any uncorrected error in the sole remaining gene can have a dramatic effect.  Loss of one gene, then damage to the remaining gene is sometimes called a “2nd hit”.  The error can be global (whole body and detectable with genetic testing), or it can be local (limited to one small region of the body, or one region of the brain). When it is local, it is undetectable and becomes an unexplained difference.
  2. Imprinting.  Imprinting is when one of the two inherited genes is silenced (turned off).  Angelman syndrome is an intellectual disability syndrome with a number of similarities to 22q13 deletion syndrome. It is caused by imprinting that turns off an important gene. Not much is known about how imprinting and chromosomal deletions interact, but obviously it would be a problem if the only remaining copy of a gene was inactivated through imprinting.  It would be another unexplained difference.
  3. Impact of gene mutation.  A gene that is completely deleted is often less damaging than a gene that has mutated.  That this the story with cancer cells.  We would much rather the cancerous cells die than grow with mutated genes.  If a chromosomal deletion includes only part of a gene, that gene can start creating proteins that actually interfere with normal cell operation.  (See When missing a gene is a good thing.)  So, a smaller deletion might disrupt part of a gene.  That partial deletion might be big trouble.
  4. Gene combinations.  It is no surprise that single gene mutations/deletions are easier to understand than 22q13 deletion syndrome, where many genes are lost.  A lot can happen when multiple genes are involved.  The best studied cases are cancer. Cancer occurs when the right (rather, wrong) combination of genes are deleted or mutated.  Recent work on autism spectrum disorder and schizophrenia show that these disorders are most often caused when a large number of gene errors add up.  Each error contributes in a small way.  In some cases, there are a few important genes, but they have little or no impact unless many other genes are also involved. These gene combinations are so subtle and poorly understood, that terms like “genetic background” act as placeholders until we understand more.  My cousin’s child and David have practically identical copies of chromosomes 22. The main genetic difference between them come from my cousin’s husband and my wife. Each spouse contributed slight, but very important, differences in their “background” genetics.  There is a lot that could be learned about 22q13 deletion syndrome genes from studying families like ours. Many other families could benefit and we could get a clearer picture of the causes of phenotype variations in 22q13 deletion syndrome.  So far, no one has asked scientists to embark on such a study.
  5. Mosaicism and somatic mutations. Recent evidence shows that a genetic error can occur during development in only one small region of the brain.  That is, some, perhaps many people have gene mutations that impact only certain areas of the brain.  These events might explain many individual variations, including things like learning disabilities.  In the case of 22q13 deletion syndrome, these silent mutations are likely to have a much more serious effect as the added mutation may interact with one or more of the 30 to 100+ missing genes.  In cases of SHANK3 mutation syndrome, the impact of SHANK3 may be greatly amplified by other lost genes in specific brain regions.  Blood tests on these patients may show up as SHANK3 syndrome, but deep in the brain they may have multiple genes lost.
  6. Genetic regulators (elements).  Since 2009 the ENCODE genetics project ( has sought to find the bits and pieces of DNA that regulate genes.  Genes make up the smallest part of DNA.  Most of DNA is made up of gene regulators.  This is very easy to understand when you realize that skin cells, brain cells, intestine cells and liver cells all have exactly the same genes.  The difference is which genes are turned on and which are turned off.  Skin cells know they are skin cells and only use skin cell genes.  The DNA is regulated in each tissue to match the genetic signature necessary to make that tissue.  Chromosome 22 deletions not only knock out genes, they knock out genetic regulators. A 4.7 Mbase terminal deletion may not hit any more genes than a 4.8 Mbase deletion, but it may hit crucial regulator sites.  It is even possible to hit the regulator site of a missing gene.  That site may impact the remaining gene on the unaffected chromosome. Even when researchers do a whole exome sequencing, they miss most of the regulators.  Gene regulators are not impossible to detect, but they can be very difficult to notice.  They are a likely cause of many unexplained differences.

Given the complexity and many opportunities for unexplained variation, we can begin to appreciate that knowing an individual’s deletion size does not provide all the answers.  However, thanks to modern tools, there are ways to study the effects of deletion size even with such wide variability.  These tools can be used to tease out which genes contribute to each medical problem.  It requires a serious commitment by parents to push researchers and medical staff toward taking full advantage of genetic reports.  Too much focus on one favorite gene hampers scientific and medical advancement. Those working on another chromosome deletion syndrome (18q deletion syndrome) have studied their syndrome wisely over the past 50 years. They are turning science into medicine for the suffers of 18q (See Making chromosome abnormalities treatable conditions).  The 18q people have developed a road map that the 22q13 deletion syndrome people can easily follow (See Consequences of chromosome 18q deletions). My past blogs have worked hard to make this point, but nothing is more convincing than seeing others take the lead with such clarity and commitment.  Why have we not benefited? The only explanation I can find is that the 22q13 deletion syndrome community lacks qualified, unbiased science leadership.  It is a fairly obvious problem, with very sad consequences.  There are no more treatments for David today than there were 30 years ago. We know which genes are missing and for many of them, we know what they do (See How do we know which genes are important and 22q13 deletion syndrome: the hope of precision medicine).  What we don’t seem to know is how to make science work for the benefit of our families.



Previous posts:

22q13 deletion syndrome: the hope of precision medicine
How do we know which genes are important
22q13 Deletion Syndrome: hypotonia
Understanding gene size
Gene deletions versus mutations: sometimes missing a gene is better.
Is 22q13 deletion syndrome a ciliopathy?
Understanding translocations in 22q13 deletion syndrome: genetics and evolution
Understanding deletion size
Can 22q13 deletion syndrome cause ulcerative colitis?
Can 22q13 deletion syndrome cause cancer?
22q13 deletion syndrome – an introduction

22q13 deletion syndrome: the hope of precision medicine

David, the backseat driver
Although non-verbal, David is clearly in charge of this trip.

David does not talk, but he does know how to express himself. In this photograph we are taking a ride to his brother’s apartment.  As soon as I arrived at David’s house, he grabbed my hand and walked me back to the car.  He pulled on the door leading to the back seat.  “Take me for a ride!  The usual place, of course!” He communicates very well considering all his disabilities, but I would love to have a medication to help him talk, or walk better, or toilet easier, or not overheat in the summer.  In fact, what I really want is a custom pill made for David.  Different patients with different size deletions have different needs. Although intellectual disability affects 100% of our kids and ASD affects up to 30%, the reality is that our children can have many different problems.  Except for a few confusing cases, kids with larger deletions have more problems and often more severe problems.

If you read my earlier blog on deletion size (Understanding deletion size), you will know that over 95% of patients with 22q13 deletion syndrome are missing from 10 to over 100 genes.  The genes near the end of the chromosome are the first ones to be deleted by a terminal deletion (the most common type).  These genes are tightly packed together.  In this region, you cannot simply say “a small deletion”.  You must know the exact deletion size to know how many genes are affected.

Precision medicine

According to the National Institutes of Health, “precision medicine” is “… an emerging approach for disease treatment and prevention that takes into account individual variability in genes, environment, and lifestyle for each person” (from: NIH Precision Medicine Initiative).  The promise of precision medicine has not reached most people because the average patient does not know which genes are most important to his/her health. For patients with 22q13 deletion syndrome, however, the genes that cause the syndrome are obviously the genes of greatest clinical importance.  The primary goal of 22q13 deletion syndrome research should be to maximize the benefit of knowing the exact genes involved on a patient-by-patient basis.  Think: “I want a pill optimized for my child”.   Of course, it is an oversimplification to think about a custom pill, but the NIH definition of precision medicine helps guide us toward more practical thinking.

In my last blog (How do we know which genes are important?) I listed the genes that are likely to contribute to hypotonia.  Categorizing the genes into  clinically meaningful categories provided us with insight into treatment.  Each gene in the list has a known effect on the brain and rest of the body.  Some genes interfere with normal brain function.  Other genes can affect peripheral myelin, a insulator that is needed to transmit information back and forth between the spinal cord and the muscles.  Still other genes can disrupt a muscle’s ability to tolerate sustained work.  Each of these categories provide important information to the physical therapist.  A child with poor sensory feedback from the muscles might be handled differently from a child with poor muscle stamina.  Precision medicine is in its infancy, used mostly in cancer treatment. However, precision medicine for 22q13 deletion syndrome could start today.  Physicians and therapists could readily benefit from a report for each person that brings together an individual’s genetics with the known functions of the missing 22q13 genes.

One might wonder how far this precision medicine idea can be taken.  Well, for next year the White House reports a 215 million dollar initiative for government supported research and promotion of precision medicine (White House Fact Sheet).  Businesses have already invested billions of dollars into electronic health records, the backbone of precision medicine. There is no question that precision medicine will bring major changes to medical practice and patient choices.

Clearing up some misconceptions

It is amusing at times to hear well-meaning parents talk about the barriers to using genetic information to guide treatment. One common misconception is that too little is known about the genes.  Actually, many of the genes have been studied for decades and the research has obvious clinical implications.  For example, at the Society for Neurosciences meeting earlier this month I talked to a young researcher from California who was working on CELSR1 (missing in about 50% of our kids).  He showed that neurons in the hippocampus essential for learning new relationships between events and places (e.g., learning to navigate a new school building or deal with a change in classroom schedule) are disrupted when CELSR1 is deleted.  What he told me next was even bigger news. A researcher in Belgium has been studying mice lacking CELSR1 for years.  It took only one email to that scientist to net a trove of information about CELSR1.  Apparently, CELSR1 is not only important for brain wiring, but also the flow of cerebrospinal fluid (CSF) in the brain. Read that: enlarged ventricles.  A radiologist who evaluates the MRI of a 22q13 deletion syndrome child will someday associate his/her findings with deletion size based on studies like these. After enough MRI reports are collected from enough patients, the association of CELSR1 with ventricle size can be confirmed.  The beauty of precision medicine is that you collect new data for the next generation each time you treat patients in this generation.  Taking your child to the doctor actually helps other patients with 22q13 deletion syndrome. Is that great, or what? For people with 22q13 deletion syndrome, it is knowing the detailed genetic information that will make it work.

Another misconception is that there is no clear relationship between deletion size and the severity of 22q13 deletion syndrome.  Actually, even if there was no clear relationship, it would still be of great value to use our knowledge of which genes are involved in each person.  But, we are sometimes faced with the confusing observation that a few kids with big deletions are more functional than others with smaller deletions.  These apparent exceptions to the rule are examples of how genetics can fool us.  Let’s use two examples to show how important knowing the basics can be.  Reading the scientific literature you can find one or two kids with tiny SHANK3 mutations/microdeletions who are more affected than one or two other kids missing a whole group of genes.  As I discussed in my earlier blog (Sometimes missing a gene is better) a gene mutation is often more damaging than deleting that gene.  Such is the case for specific mutations of SHANK3, ATXN10, CELSR1 and other genes on 22q13.  That is part of the reason I use the term “22q13 deletion syndrome”, which distinguishes deletions from mutations.  The second example is the clumping of important genes on the distal part of the chromosome. Because the genes are not evenly distributed on the chromosome, someone with a 1.5 Mbase deletion and someone with a 2.5 Mbase are actually missing the same genes.  Deletion size is not a measure of gene loss.  It simply provides a map to the list of genes that are deleted.  Comparisons have to be made after making a list of genes.

There are other reasons for a conflict between deletion size and severity of 22q13 deletion syndrome.  One recent study has shown that de novo chromosomal deletions (the most common type) often include mutations and other deletions elsewhere on the chromosome or on other chromosomes.  This more widespread occurrence of genetic errors does not tend to show up in the parents or siblings of a child with a de novo deletion.  That is, a diagnosis is 22q13 deletion syndrome raises the possibility that there are more genetic errors elsewhere in the DNA.  Precision medicine will someday not only include the deletion size, but a list of other genes that show potential issues.  There are other reasons for the unusual cases that I won’t go into, but larger deletions affect more genes and generally cause more problems. Of course, individual differences do matter.  That is why it is called precision medicine.

The future is now

My posting on hypotonia landed me an opportunity to give a guest lecture to a graduate physical therapy class.  The lecture was on the genetics of infant hypotonia.  I ended the lecture with a “hopeful warning” that all of medicine is about to change.  It was a warning, because all clinical practitioners will need to understand the implications of genetics in their practice, and it was hopeful because the lives of patients are about to get better.  It may be a while before we can go to an apothecary for a customized pill, but we can reap benefits today.  Your physicians, nurses and therapists could begin receiving guidance curated from the currently available literature on genes.  Of course, someone has to compile the information.  Perhaps we need to convene a conference that brings together experts on each gene with medical practitioners who would use the information. I have seen a number of conferences for 22q13 deletion syndrome, but none like that.

I should probably get a detailed genetic report for David and combine that with my own readings on his genes so that he can benefit from the promise of precision medicine. I am torn by a moral dilemma. I don’t want to be biased in my pursuit of 22q13 genetics. Whether we like it or not, we are always biased by what our own child needs. Not knowing David’s details is, in a way, liberating. I am hanging out with David as I write this. We are watching the Graceland video with Paul Simon.  If you know the history behind that video, it is a reminder that everyone matters, regardless of their skin color, which is to say, regardless of their genetics.



Previous posts:

How do we know which genes are important
22q13 Deletion Syndrome: hypotonia
Understanding gene size
Gene deletions versus mutations: sometimes missing a gene is better.
Is 22q13 deletion syndrome a ciliopathy?
Understanding translocations in 22q13 deletion syndrome: genetics and evolution
Understanding deletion size
Can 22q13 deletion syndrome cause ulcerative colitis?
Can 22q13 deletion syndrome cause cancer?
22q13 deletion syndrome – an introduction

Understanding gene size

David fathers day 2015Father’s Day gifts from David

David has a terminal deletion of chromosome 22 caused by a balanced translocation.  Like nearly everyone with 22q13 deletion syndrome, he is missing a lot more than one gene.  What, exactly, does that mean?

DNA and genes

Each gene is made up of many “bases”.  DNA has two strands (strings) that grip each other tightly. Imagine a bunch of magnets threaded onto a string like pearls. Now make two of these in your mind and hold them near each other.  Slowly bring them close together.  When they get near, the north poles of magnets from one string will start to find the south poles from the other string. When the magnets come together, opposite poles will grab each other.  Anywhere north faces north, or south faces south, that pair will repel each other until one flips around and the opposites unite.  DNA is made of chemical strings that have opposite poles. These opposites find their mate and the two DNA strands lock together. Each time a north meets a south you get a “base pair”.

Magnets can only make one type of partnership (north attracted to south).  DNA actually has two kinds of partnerships from four chemical bases.  The bases are abbreviate T, A, G and C.  T and A attract each other.  G and C attract each other.  If you make a string like this: -T-A-G-G-C-A-, the matching string will always look like this: -A-T-C-C-G-.  That is, the strings stick to each other in this way:


Voilà! You have a small strand of DNA.  This miniature DNA has 6 base pairs.  The order of the base pairs describe the protein that this segment of DNA makes.  The lower strand is kind of mirror of the upper strand. If you know what is on one strand you can always figure out the other strand. Thus, we now know a bunch of properties of DNA:  1) The sequence of base pairs describes how to make a protein, 2) DNA is strongly stuck to itself, 3) DNA keeps a mirror copy of itself available at all times, and 4) the length of the DNA can be measured by counting the number of base pairs.  There is a lot more to learn about DNA, but this is enough to discuss gene size.

Big genes are easier to find

In a previous posting I explained that 95% of all people with 22q13 deletion syndrome are missing at least 1 Mbase from their chromosome (see Understanding deletion size).  1 Mbase means 1,000,000 (1 million) base pairs along the two parallel strands of DNA.  Genes are segments of the long strings, like chapters in a book.  And, like many books, some chapters are long and some are short.  There are 32 genes in the distal 1 Mbase of 22q13, many of which influence brain function. Chromosome deletion syndromes are inherently difficult to study because so many genes are involved.  It is hard enough to study and understand the impact of losing a single gene.  It is much harder to study and understand 22q13 deletion syndrome, where many genes are missing.

This problem with studying multiple genes is not unique to 22q13 deletion syndrome.  It shows up in neuropsychiatric disorders like autism and schizophrenia, each of which have hundreds of associated “risk factor” genes.  Autism, for example, results from various combinations of these many genes (see review by Gratten et al., 2014).  Chromosomal deletions are known to operate in a similar way (see contiguous gene syndrome). Each missing gene weakens the normal operation of the brain.  No one gene needs to be “dominant” for the combined loss to be devastating, especially when so many brain-related genes are missing at once.

Not everyone thinks of 22q13 deletion syndrome this way.  Much of the current thinking about the genes lost in 22q13 deletion syndrome focuses on one or two genes that code for synaptic proteins.  The term “synaptopathy” has been used a lot recently, but that word originates from the study of the inner ear where they are able to clearly demonstrate the relationship between synaptic function and hearing loss (Sergeyenko et al., 2013).  The same does not hold true for 22q13 deletion syndrome. Synapses are involved, but the synapse may not be the primary site of dysfunction (see Is 22q13 deletion syndrome a ciliopathy?). For many years no one thought primary cilia were important. Now, ciliopathies are a recognized type of brain dysfunction despite the fact that synapses are also involved. Science often goes off in a wrong direction; it is part of the process.

There is another reason that synaptic genes have taken the spotlight.  The synaptic genes of 22q13 are relatively large genes.  Defects of these genes are simply easier to notice.  If we look at the history of 22q13 deletion syndrome, the first cases were discovered in people with very large deletions and with the most “severe” phenotype.  As the research in 22q13 deletion syndrome advanced, smaller and smaller deletions were identified and studied.  At the moment, the only gene getting any attention is a large gene that has a large effect when mutated, even though mutations do not necessarily tell you what happens when a gene is deleted (see When missing a gene is a good thing).  So, why does size matter?

Pie chart of mRNA size of first 1 mbase
Genes lost in a 1 Mbase deletion of 22q13 sorted by their sizes (mRNA size).
Right click on the graph to see a full size image.

The pie chart shows the 32 genes missing in 95% of patients with 22q13 deletion syndrome.  They are in order of size. The largest gene is SBF1 and the second largest is SHANK3.  The genes continue in descending order of size in a counter-clockwise direction.  Although the reality is a bit more complex, it is generally true that the likelihood of a gene mutation depends on the gene’s size.  This pie graph shows that the 10 largest genes account for half of the “protein-coding” DNA in the first 1 Mbase.  To put it another way, you are twice as likely to incur a mutation of SHANK3 than incur a mutation of MAPK8IP2, simply because SHANK3 is twice as large. SHANK3 is 16 times larger than SYCE3.  So, when studying mutations, SHANK3 can show up more often simply because it is big.

As I noted above, no one knows what a complete deletion of SHANK3 might do on its own. A gene can have a severe phenotype when mutated, but might do little or no harm when missing altogether.  SHANK3 may have some contribution to 22q13 deletion syndrome, but its relative contribution is very poorly understood.  There are other 22q13 genes that have severe consequences after mutation, usually when both copies are mutated.  We have discussed some of these previously (Can 22q13 deletion syndrome cause cancer?, Can 22q13 deletion syndrome cause ulcerative colitis? and Is 22q13 deletion syndrome a ciliopathy?).  Another gene is SBF1, which causes Charcot-Marie-Tooth disease type 4B3. The phenotype includes intellectual disability.  MAPK11 and MAPK12 are involved in responses to oxidative stress, and are likely important to recovery from infection and brain trauma. SBF1 is large, but MAPK11 is much smaller. SCO2 is one of the smallest genes, yet it is implicated in a series of severe, including fatal, syndromes (DiMauro et al., 2012).  What happens when all of these genes are deleted together?  You get 22q13 deletion syndrome.

The take-home message is that certain genes are more likely to come under the microscope (literally and figuratively) simply because they are larger genes. Being large makes a gene easier to study (usually), but it does not necessarily confer importance. When a gene gets popularized in the scientific literature, lots of papers are published on that one gene, at least for a while.  Scientists will focus on genes that get them grants and publications. That is how science typically works, even if it is not necessarily the best approach to finding effective treatments that families really need. The direction of science can be influenced by patient groups, but choosing the right direction requires a deep understanding of the science (the current state of research), science (the discipline) and scientists (who do science).


Previous posts:
Gene deletions versus mutations: sometimes missing a gene is better.
Is 22q13 deletion syndrome a ciliopathy?
Understanding deletion size
Can 22q13 deletion syndrome cause ulcerative colitis?
Can 22q13 deletion syndrome cause cancer?
22q13 deletion syndrome – an introduction