Lecture 32: Genetic Screening PDF

Lecture 32: Genetic Screening Robin T. Varghese Ph.D. [email protected] Original presentation content credit: Pawel Michalak, PhD Learning Objectives 1. Distinguish between genetic testing versus genetic screening. 2. Recognize criteria necessary for an ethical and successful screening program....

Lecture 32: Genetic Screening Robin T. Varghese Ph.D. [email protected] Original presentation content credit: Pawel Michalak, PhD Learning Objectives 1. Distinguish between genetic testing versus genetic screening. 2. Recognize criteria necessary for an ethical and successful screening program. 3. Identify the genetic disorder that is nationally screened for in infants. 4. Identify other disorders commonly screened for in infants. 5. Identify the one genetic disorder that has been strongly pushed for an adult screening program and compare and contrast the arguments for and against such a program. 6. Identify examples of both successful and non-successful heterozygote screening programs. 7. Identify uses for GWAS for assigning risk to chromosome alleles and disease. Genetic Screening Population-based method for identifying persons with increased susceptibility for a genetic disease regardless of family history and independent of clinical status Genetic Screening vs. Genetic diagnostic Testing • Genetic Screening-determine which individuals may have a higher risk factor for the disease • Detect treatable human diseases in their pre-symptomatic stage • Does not determine a diagnosis • Identifies a subset whom further, more exact, diagnostic tests should be carried out • Genetic (diagnostic) testing – analysis of chromosomes, DNA, RNA, proteins, or other analytes, to detect abnormalities that can cause a genetic disease • Indications include prenatal diagnosis, heterozygote carrier detection, and pre-symptomatic diagnosis of genetic disease Criteria for effective genetic screening • Treatment is available. • A rapid and economic laboratory test is available that detects the appropriate metabolite or gene product. • Early institution of treatment, before symptoms become manifest, reduces or prevents severe illness. • Routine observation and physical examination will not reveal the disorder in the newborn—a test is required. • The condition is frequent and serious enough to justify the expense of screening; that is, screening is cost-effective. • The laboratory test is highly sensitive (few false-negatives) and reasonably specific (few false-positives). Sensitivity and Specificity What is it: Sensitivity and specificity are two important parameters used to assess the performance of a genetic screens or diagnostic tests How is it evaluated: To test the sensitivity (true positive rate) or specificity (true negative rate) of your test/screen, you need a group of individuals who are: -known to have the condition or disease ("gold standard") & -known NOT to have the condition or disease (control) What is determined: You perform the test on this group and calculate the true positive(TP) rate and true negative(TN) rate. Why?: typically for the evaluation of a new diagnostic test or screening tool. It's crucial to assess the performance of the test before it is widely used in clinical practice Sensitivity: ability for a screen/test to identify affected individuals. • • Measured by the proportion of true positives (TP) E.g., if a test only identifies 50 affected individuals out of 100 individuals who are truly affected with the disease, then the sensitivity for that test would be 50% Specificity: ability for a screen/test to identify unaffected individuals • • (TRUTH) Measured by the proportion of true negatives (TN) E.g., if a test only identifies 80 unaffected individuals out of 100 individuals who do not have the disease, then the specificity for that test would be 80% What test predicts N=200 (All with Disease) Healthy (NO disease) https://step1.medbullets.com/stats/101006/testing-and-screening Sensitivity Specificity & https://step1.medbullets.com/stats/101006/testing-and-screening Types of Genetic Screening • Screening before birth: • Fetal cell screening • Ultrasound • Maternal serum • Screening after birth: • Newborn screening • Carrier (heterozygote) • Adult pre-symptomatic screening Newborn screening Condition Frequency (per 100,000 newborns) Congenital hearing loss 200 Sickle cell disease * 47 Hypothyroidism * 28 Phenylketonuria * 3 Congenital adrenal hyperplasia 2 Severe combined immunodeficiency 2 Galactosemia * 2 Maple syrup urine disease ≤1 Homocystinuria Biotinidase deficiency Identifies pre-symptomatic infants ≤1 ≤1 * Tested in all states Newborn Screening in VA • • • • • • • • • • • • • • • • • 3 hydroxy 3 methylglutaryl-CoA lyase deficiency 3-methylcrotonyl-CoA carboxylase deficiency Argininosuccinic acidemia Beta ketothiolase deficiency Biotinidase deficiency Carnitine uptake deficiency Citrullinemia Congenital adrenal hyperplasia Congenital hypothyroidism Critical congenital heart disease Cystic fibrosis Galactosemia Glutaric acidemia type I Hearing loss Homocystinuria Isovaleric acidemia Long chain hydroxy acyl-CoA dehydrogenase deficiency • • • • • • • • • • • • • • • • • • Maple syrup urine disease Medium chain acyl-CoA dehydrogenase deficiency Methylmalonyl adenosyl- cabalamine synthesis defects Methylmalonyl-CoA mutase deficiency Mucopolysaccharidosis type-I Multiple CoA carboxylase deficiency Phenylketonuria Pompe disease Proprionic acidemia Severe combined immunodeficiency Sickle beta thalassemia Sickle cell anemia Sickle hemoglobin C disease Spinal muscular atrophy Trifunctional protein deficiency Tyrosinemia type I Very long chain acyl-CoA dehydrogenase deficiency X-linked adrenoleukodystrophy The commonwealth of Virginia screens newborns for 35 diseases https://newbornscreening.hrsa.gov/your-state/virginia Tandem Mass Spectrometry • Simultaneous detection of dozens of biochemical disorders • Can be utilized to analyze metabolite levels • 1 sample can detect the elevated metabolite with fewer false positives and detect other disease simultaneously • e.g. PKU, galactosemia Phenylketonuria (PKU)- Newborn Deficiency in enzymes that break down phenylalanine Screening phenylpyruvic acid PAH gene mutation proteins dopamine melanin Symptoms: include mental retardation, stunted growth, eczema, small head size, seizures Biochemistry of phenylketonuria. Phenylketonuria (PKU)- Newborn Screening Galactosemia- Newborn Screening • Deficiency in enzymes that break down galactose (often GALT mutations) • Symptoms: Feeding difficulties, lethargy, failure to thrive, jaundice, brain damage, seizures cataracts Heterozygote testing • High frequency of carriers, at least in a specific population • Availability of an inexpensive and dependable test with very low false-negative and false-positive rates • Access to genetic counseling for couples identified as heterozygotes • Availability of prenatal diagnosis • Acceptance and voluntary participation by the population targeted for screening Heterozygote (Carrier) testing Tay-Sachs • 1 in 30 carrier screening among Ashkenazi Jewish populations • Carrier screening determines if an individual carries one disease-causing copy of the HEXA gene. • As a result of screening, TaySachs declined by 90% Heterozygote (Carrier) testing Cystic Fibrosis Heterozygote (Carrier) Cystic fibrosis (CF) is an autosomal recessive condition characterized by viscous testing mucus in the lungs along with involvement of the digestive system and sweat glands. CF is caused by mutations in the CFTR gene. Most cystic fibrosis cases in the U.S. are caused by a mutation called delta F508 (∆F508). American College of Medical Genetics (ACMG) originally recommended a 21mutation panel in 2001. CFTR gene > 1700 mutations discovered: Now the ACMG has a 215 variant expanded panel to detect more variants. ACMG, The American College of Medical Genetics and Genomics ACMG Screening Panel- CF Ethnic Group Incidence of Cystic Fibrosis Caucasian 1 in 3,200 Carrier Detection Rate Probability of expanded without Testing ACMG CF panel (215 variants) 1/25 89.4% African 1 in 15,300 American Hispanic 1 in 9,500 American Asian American 1 in 32,100 1/65 65.6% 1/46 74.8% 1/90 54.5% Ashkenazi Jewish 1/25 94% 1 in 3,300 Sensitivity < 100% Limitations of Genetic Screening • Never 100% accurate • Mosaicism • Human Error • May not detect all disease-causing mutations • Multiple mutations in the same gene may lead to the same disease- allelic heterogeneity • Other considerations • Anxiety in carriers • Lack of curative treatments • Ethical- discrimination by insurance companies • Polygenic diseases are more common yet more difficult to screen Disease Associations: Genome wide association studies (GWAS) Goal: identify genes associated with a particular disease (or another trait) which is known to be heritable e.g., height, Schizophrenia, etc. • Specific aims are distinct: 1. Identify statistical connections between specific locations in the genome and the phenotype 2. Generate insights on genetic architecture of phenotype • -Many small genetic effects dispersed across the genome • -Few large effects concentrated in one area 3. Build statistical models to predict phenotype from genotype • “Show me your genome and I will tell you what diseases you will get” GWAS Methodology Case vs. Control 1. Find thousands of people who differ for the disease of interest: • Two groups of participants: people with the disease vs. similar people without the disease 2. Isolate DNA • Utilize DNA Microarray to identify strategically selected markers of genetic variation • SNPS: usually 105 – 106 (SNPs) • SNPs that are (statistically) significantly more frequent in people with the disease are said to be ‘associated’ with the disease • These SNPs maybe in or near genes that may be important to disease Single Nucleotide Polymorphisms (SNPs) are common simple DNA variants which can be used as biological markers, helping scientists locate genes that are associated with disease. Can you find the associated SNP? Cases: AGAGCAGTCGACAGGTATAGCCTACATGAGATCGACATGAGATCGGTAGAGCCGTGAGATCGACATGATAGCC AGAGCCGTCGACATGTATAGTCTACATGAGATCGACATGAGATCGGTAGAGCAGTGAGATCGACATGATAGTC AGAGCAGTCGACAGGTATAGTCTACATGAGATCGACATGAGATCGGTAGAGCCGTGAGATCGACATGATAGCC AGAGCAGTCGACAGGTATAGCCTACATGAGATCAACATGAGATCGGTAGAGCAGTGAGATCGACATGATAGCC AGAGCCGTCGACATGTATAGCCTACATGAGATCGACATGAGATCGGTAGAGCCGTGAGATCAACATGATAGCC AGAGCCGTCGACATGTATAGCCTACATGAGATCGACATGAGATCGGTAGAGCAGTGAGATCAACATGATAGCC AGAGCCGTCGACAGGTATAGCCTACATGAGATCGACATGAGATCGGTAGAGCAGTGAGATCAACATGATAGTC AGAGCAGTCGACAGGTATAGCCTACATGAGATCGACATGAGATCTGTAGAGCCGTGAGATCGACATGATAGCC Controls: Associated SNP AGAGCAGTCGACATGTATAGTCTACATGAGATCGACATGAGATCGGTAGAGCAGTGAGATCAACATGATAGCC AGAGCAGTCGACATGTATAGTCTACATGAGATCAACATGAGATCTGTAGAGCCGTGAGATCGACATGATAGCC AGAGCAGTCGACATGTATAGCCTACATGAGATCGACATGAGATCTGTAGAGCCGTGAGATCAACATGATAGCC AGAGCCGTCGACAGGTATAGCCTACATGAGATCGACATGAGATCTGTAGAGCCGTGAGATCGACATGATAGTC AGAGCCGTCGACAGGTATAGTCTACATGAGATCGACATGAGATCTGTAGAGCCGTGAGATCAACATGATAGCC AGAGCAGTCGACAGGTATAGTCTACATGAGATCGACATGAGATCTGTAGAGCAGTGAGATCGACATGATAGCC AGAGCCGTCGACAGGTATAGCCTACATGAGATCGACATGAGATCTGTAGAGCCGTGAGATCGACATGATAGCC AGAGCCGTCGACAGGTATAGTCTACATGAGATCAACATGAGATCTGTAGAGCAGTGAGATCGACATGATAGTC This approach generates (~ 105 – 106 ) total hypotheses tests and p values (p-value measures the probability of obtaining the observed results, assuming that the null hypothesis is true) SNPs indicated in red font Manhattan Plot of GWAS Declaring only one association in Chr7 What happens if we use a p-value threshold of α=0.05 (black line) to declare results as significant? SNPs with big difference between cases & controls α= 5 x 10-8 Or a=.00000005 1 test= 5% chance of obtaining our observed test statistic or a more extreme statistic. 5/100 false discoveries: 5% false positive α=0.05 Solution: be very selective in what results we declare as significant. Currently the standard is 5x10-8 Each colored dot represents the pvalue from hypothesis testing in a single SNP SNPs without major difference between cases and controls Results of famous WTCCC study of seven diseases on 14,000 cases and 3,000 shared controls (Nature, 2007) Total found: 13 significant findings at level 5*10-8 Wellcome Trust Case Control Consortium (WTCCC) Manhattan plot of Alzheimer’s Disease α= 5 x 10-8 Our GWAS findings do not explain heritability • Height: • From twins and family study, about 80% of height variability is heritable • Huge height GWAS (n>40K ) found SNPs explaining ~10% of height variability • Diseases: Schizophrenia, heart disease, cancers,… • Heritability: ~80% • For none of these, GWAS gives more than 5%-10% • Basically, for all complex traits investigated a major gap remains! Where is the missing heritability? Theories: 1. Rare variants not covered by GWAS 2. Complex associations: combinations of multiple SNPs and environment 3. Lack of power: the effects are weak, we need much more data 4. Epigenetic effects: heritability influenced by transgenerational epigenetics 5. Population stratification- a disease may occur in certain populations more often than other populations and these populations may have more common alleles in general (not associated with the disease in question). 6. Multiple hypothesis testing- the more hypotheses (SNPs) you test the chance of finding association by chance alone increases (false discoveries) References The hyperlinks embedded within the lecture notes Word document provide ample references for this material.