Genome-Wide Association Study - An Explanation for Beginners

Summary

In this video, Nuna Crevallo explains the concept of a Genome-Wide Association Study (GWAS). She describes how GWAS identifies associations between genetic variations, known as single nucleotide polymorphisms (SNPs), and specific traits like BMI. Crevallo outlines the process, which involves large-scale data collection, genotyping, regression analysis, and the use of statistical methods to determine significant genetic associations. She also explains the importance of correcting for false positives and how the findings can lead to better understanding of genetic diseases and traits, potentially improving treatments and prevention.

Takeaways

• 🧬 Genome-wide association studies (GWAS) investigate associations between genetic variations and specific physical traits.
• 🧪 Single nucleotide polymorphisms (SNPs) are key genetic variations, where different nucleotides occur at the same position in the DNA.
• 👥 Humans share 99.9% of their genome, but the 0.1% of SNPs contribute to individual differences.
• 🌐 GWAS requires large sample sizes, typically from thousands to hundreds of thousands of people, to detect genetic associations.
• 🔄 The study focuses on a specific trait or phenotype, such as BMI, and analyzes genetic data across millions of SNPs.
• 📊 A regression analysis is performed to see if the number of SNP variants (alleles) is linked to the physical trait being studied.
• 📉 The strength of association is measured by p-values, which indicate the likelihood that an association is due to chance.
• 🔢 A Bonferroni correction is applied to adjust for false positives when testing millions of SNPs, using a stricter significance threshold.
• 🏙 A Manhattan plot visually represents significant associations, with SNPs positioned along chromosomes and their p-values shown.
• 🔍 Identifying SNPs associated with traits can lead to discoveries of important biological mechanisms and improve disease treatment.

Q & A

What is a Genome-Wide Association Study (GWAS)?

A Genome-Wide Association Study (GWAS) is a method used to discover associations between variations in our genetic code and certain physical traits.

What are single nucleotide polymorphisms (SNPs)?

SNPs are variations at a single position in a DNA sequence among individuals. They are a key element in understanding the genetic causes for human traits.

Why is it important to understand the genetic makeup of humans?

Understanding the genetic makeup is important to determine the extent to which certain traits are genetic and to identify biological mechanisms that affect those traits.

How does a GWAS help in understanding traits like BMI?

A GWAS helps in understanding traits like BMI by looking for variance in DNA that might impact it, such as genes that affect metabolism.

What is the significance of having a large sample size in a GWAS?

A large sample size, ideally in the thousands or hundreds of thousands, is significant because it helps to minimize confounding genetic variation and increases the statistical power of the study.

What does genotyping involve in the context of a GWAS?

In a GWAS, genotyping involves recording the nucleotides of each person at many known SNP locations, often over 2 million SNPs.

How is the association between genetic data and a phenotype computed?

The association between genetic data and a phenotype is computed using a genome-wide association analysis program, which performs regression analysis for every SNP in the dataset.

What role does the software PLINK play in GWAS?

PLINK is a commonly used software for GWAS that performs quality control filters on the genetic dataset and carries out the actual association analysis.

What is a p-value in the context of GWAS, and why is it important?

A p-value measures the likelihood that the association found in the distribution of data points was due to random chance. It is important for determining the statistical significance of the association between a SNP and a phenotype.

What is a Bonferroni correction, and why is it used in GWAS?

A Bonferroni correction is a statistical method used to adjust the threshold for significance when dealing with multiple comparisons, like in GWAS, to reduce the chance of false positives.

How are the results of a GWAS typically visualized?

The results of a GWAS are typically visualized using a Manhattan plot, which plots each SNP's position along the chromosome and its associated p-value after a negative log transformation.

What is linkage disequilibrium, and how does it relate to GWAS?

Linkage disequilibrium refers to the non-random association of alleles at different loci. In GWAS, significant SNPs often cluster together due to linkage disequilibrium, which can indicate the presence of important genes.

Outlines

00:00

🧬 Understanding Genome-Wide Association Studies

Nuna Crevallo introduces Genome-Wide Association Studies (GWAS), explaining that they are used to find links between genetic variations and physical traits. She starts by defining the human genome and its composition of nucleotides (adenine, thymine, guanine, and cytosine). She points out that only 0.1% of our DNA varies among individuals, which accounts for human diversity. Crevallo then describes Single Nucleotide Polymorphisms (SNPs) as locations in DNA where different nucleotides can occur in a significant portion of the population. These SNPs are crucial for understanding the genetic basis of traits. The process of GWAS involves collecting genetic data from a large group of people, genotyping them for millions of SNPs, and correlating this data with a trait of interest, such as BMI. Crevallo explains that association analysis is performed using software like PLINK, which conducts regression analysis for each SNP to determine if there's a relationship with the trait. The effectiveness of the predictive line is measured by the p-value, which indicates the likelihood that the observed association is not due to chance. The analysis is computationally intensive but can be expedited with efficient processing techniques.

05:02

📊 Analyzing GWAS Results and Significance

In the second paragraph, Crevallo discusses the analysis of GWAS results, focusing on p-values as a measure of statistical significance. She clarifies that a p-value threshold of 0.05 is commonly used to determine if an association is significant and not due to chance. However, with millions of SNPs tested, the likelihood of false positives increases, leading to the use of a Bonferroni correction to adjust the threshold for significance. The field has adopted a threshold of 5x10^-8 to reduce false positives. Crevallo then explains how Manhattan plots are used to visualize the results of GWAS, plotting each SNP's chromosome position against its p-value after a negative log transformation. Significant SNPs are indicated by dots above a certain line on the plot. She mentions that clusters of significant SNPs may be due to linkage disequilibrium and can help identify gene locations associated with the phenotype of interest. Finally, Crevallo suggests that understanding these genetic associations can lead to improved prevention and treatment of genetic diseases, potentially saving many lives. She concludes by inviting suggestions for future topics and wishing the viewers a good day.

Keywords

💡Genome

A genome is the complete set of genetic material present in an organism. In the video, it refers to all of the DNA contained in human cells, which includes 23 pairs of chromosomes. Understanding the genome is fundamental to identifying variations in DNA that affect physical traits.

💡Single Nucleotide Polymorphisms (SNPs)

SNPs are variations at a single position in the DNA sequence among individuals. In the video, SNPs are highlighted as key markers in the study of genetic traits, such as BMI, as they help researchers pinpoint genetic differences that could influence specific traits.

💡Genome-Wide Association Study (GWAS)

A GWAS is a study that examines the entire genome of multiple individuals to find genetic variations associated with a particular trait. In the video, the term refers to the method used to discover connections between genetic differences (SNPs) and physical traits like BMI.

💡Phenotype

A phenotype is the observable physical trait or characteristic of an organism, such as eye color or BMI. The video discusses how phenotypes are analyzed in a GWAS to find genetic links to these traits by comparing them with SNP data.

💡P-value

A p-value is a statistical measure used to determine the likelihood that an observed association between SNPs and a trait is due to chance. In the video, p-values are used to assess the significance of the results from the GWAS, with smaller p-values indicating stronger evidence for a true genetic association.

💡Bonferroni Correction

The Bonferroni correction is a statistical method used to adjust the threshold for significance in studies with multiple comparisons to reduce false positives. In the video, this correction is applied to handle the large number of SNPs analyzed in GWAS, ensuring more reliable results.

💡Linkage Disequilibrium

Linkage disequilibrium refers to the non-random association of alleles at different loci in a population. While not explained in depth in the video, it is mentioned as a factor influencing the clustering of significant SNPs in genetic studies, helping researchers identify which genes are associated with traits.

💡Regression Analysis

A regression analysis is a statistical process used to determine the relationship between variables. In GWAS, as explained in the video, it is used to analyze the relationship between SNPs (independent variables) and a trait like BMI (dependent variable), predicting how changes in SNPs affect the trait.

💡Effect Size

Effect size refers to the magnitude of the relationship between a SNP and a trait in a GWAS. In the video, effect size is discussed in relation to the slope of the regression line, indicating the strength of the association between SNP variants and traits like BMI.

💡Manhattan Plot

A Manhattan plot is a graphical representation of GWAS results, where each SNP is plotted based on its chromosomal location and its significance (p-value). In the video, this plot helps visualize which SNPs are significantly associated with a trait, with higher points indicating stronger associations.

Highlights

Introduction to genome-wide association study (GWAS) and its importance in discovering associations between genetic variations and physical traits.

Explanation of the human genome consisting of 23 pairs of chromosomes and made up of nucleotides: adenine (A), thymine (T), guanine (G), and cytosine (C).

Humans share 99.9% of their genetic code, with only 0.1% responsible for individual diversity.

Introduction to single nucleotide polymorphisms (SNPs), which represent variations in a single nucleotide at specific positions in the genome.

SNPs play a key role in understanding the genetic basis of heritable traits and can reveal biological mechanisms affecting traits.

Large sample sizes, typically in the hundreds of thousands, are needed for accurate GWAS results, especially for complex traits like BMI.

Ethnic homogeneity within the sample is critical to minimize confounding genetic variations.

Genotyping is the process of recording genetic information at millions of known SNPs for each individual in the study.

A genome-wide association analysis uses regression analysis to examine the relationship between SNP variants and traits such as BMI.

Regression analysis can account for the number of alleles (0, 1, or 2) for each SNP and their relationship to a phenotype like BMI.

P-values help determine the statistical significance of associations between SNPs and traits, with lower p-values indicating stronger associations.

A Bonferroni correction is used to adjust for multiple comparisons in GWAS, lowering the threshold for significance to reduce false positives.

The common threshold for GWAS significance is 5 × 10^-8, which helps minimize false positive associations.

Manhattan plots visually represent significant SNPs, with the x-axis showing SNP positions on chromosomes and the y-axis showing p-values.

Clusters of significant SNPs often indicate regions of interest that may be linked to important genes affecting the trait being studied.

Transcripts

hello i am nuna crevallo

and today i'll be explaining what a

genome-wide association study

is the quick summary is that a

genome-wide association study or a

g-loss for short

is the discovery of associations between

certain variations in our genetic code

and a certain physical trait to explain

how all of this can be done

we must begin at the start with an

explanation of single nucleotide

polymorphisms

all of human's genetic code is called

the genome

in the nucleus of each cell the genome

is split between 23 pairs of chromosomes

our dna is made up of an extremely long

chain of connected single units called

nucleotides

which come in four forms adenine

thymine guanine and cytosine

these variations are often shortened to

a t g and c

respectively incredibly humans share

99.9 percent of their genetic makeup

meaning

only 0.1 percent of our entire dna is

responsible for the diversity we see

between individuals

so let's say we choose a spot along our

dna and find that 95 percent of people

have an a nucleotide here while the

remaining five percent have a t

nucleotide

each of these forms is called a variant

given that this location on human dna

can have multiple forms

it is dubbed a single nucleotide

polymorphism or a snip

for short

these snips are key in understanding the

genetic causes for human traits

while some traits such as what la-based

nva team

prefer are almost entirely environmental

other traits like eye color

are extremely heritable snips can help

us understand to what extent certain

traits are genetic

as well as what biological mechanisms in

our body might be affecting those traits

to do this an association analysis can

be performed

let's pretend that we are trying to find

genetic associations with bmi

meaning we're trying to figure out what

variance in our dna might be impacting

bmi

such as genes that might increase or

decrease the efficiency of metabolism in

our body

we first need a large sample size of

volunteers at least in the thousands

were preferably closer to hundreds of

thousands

this cohort of people should ideally be

of identical ethnicities to minimize

confounding genetic variation

from other factors then each person

needs to be genotyped or have their

nucleotides recorded at many known snip

locations

often each person will have genetic

information recorded for over 2 million

snips

after that we need to record each

person's bmi once we have recorded the

genome and the trait or phenotype

of a large number of people an

association between the two can be

computed

the next steps involve using a

genome-wide association analysis program

a commonly used software for this is

called plink using the software

some quality control filters can be

performed on the genetic data set

removing any individuals or snips that

may have faulty information

next the actual association analysis is

done

for every snip in the data set a

regression analysis is performed with

each individual being a data point

for example let's say the program is

performing a regression analysis for

snip id number one

which has two possible alleles an a and

a t

each individual in the data set then has

the number of t alleles they have in

their genome for that snip plotted

against the physical trait of interest

in this case bmi keep in mind that human

dna contains a copy from our father

and a copy from our mother meaning that

a person's allele combinations for the

snip can be a

a a t or t t this would be coded as a 0

a 1 and a 2 respectively

once each individual is plotted the

program tries to draw a line to the data

that best predicts the relationship

between the number of alleles and the

phenotype

if there is no association between that

snip and bmi the regression line would

essentially be a horizontal line

however if an association is present you

can expect the regression line to have

some sort of slope

the effectiveness of the line at

predicting the data points determines

the p-value

the p-value is a measure of the

likelihood that the association found in

the distribution of data points was due

to random chance

given that there is no association

between the snip and bmi

this means the stronger the data points

cluster around a sloped regression line

the less likely it is that it is due to

random chance producing a small p-value

for each snip the p-value calculated is

recorded as well as the slope of the

regression line which is also known as

the effect size

this regression analysis repeats for

every single snip in the data set

when you are working with millions of

snips this will often take hours if not

days

for a computer to accomplish luckily

programs like playing can take advantage

of efficient processing such as

multi-threading

to finish the analysis quicker

furthermore these programs allow for the

addition of covariates which are other

traits that may affect the phenotype

that is of interest

so for example the amount of exercise

done weekly by a person

certainly has an effect on bmi so if

this information is present for the

people in the data set

the regression analysis can also account

for the effect of this covariate when it

competes the effect of the genotype

alone

once p-values have been calculated for

all snips how do we determine if an

association is considered significant or

not

once again this has to do with p-values

you might have heard before

of using five percent as a threshold for

statistical significance

this means that only results producing

p-values lower than .05

are considered to be due to a real

association present and not random

chance

however with a p-value of point zero

five you still have a five percent

chance of producing a false positive

meaning around five percent of your

significant results may still be because

of random chance

when you are working with millions of

snips you are likely to produce

thousands of false positives

which can muddle results and diminish

the statistical power of your study

so a bonferroni correction is performed

which transforms the threshold required

for achieving significance by taking the

typical .05 threshold

and dividing it by the number of snips

in the analysis

the quantitative genomics field has

adopted 5 times 10 to the negative 8 as

the default threshold for significance

this will sharply reduce how many snips

find a significant association with the

phenotype

but chances are that most of those snips

were false positives to begin with

to visualize these results a manhattan

plot is produced

a manhattan plot takes each snip and

plots its position along the chromosome

on the x-axis

and its associated p-value on the y-axis

the p-value undergoes a negative log

transformation to make it easier to read

essentially dots above this line

represent snips that are significantly

associated with bmi

often significant snips will be

clustered together due to linkage

disequilibrium

a topic beyond the scope of this video

but what is important is that these

clusters need to be analyzed using snip

databases that indicate what genes are

present for those regions

identifying gene locations associated

with the phenotype may uncover important

biological mechanisms that had not been

found before

with these findings the prevention and

treatment of certain genetic diseases

may improve drastically and hopefully

many lives can be saved in the future

thank you for watching this video if you

have any suggestions for a quantitative

genomics topic for me to explain

let me know in the comments i'm nuna

crevallo and i hope that you have a good

day