Please go through the below explanation for details on table and plot interpretations
Table column interpretation:
SnpEff is an open source tool that annotates variants and predicts their effects on genes by using an interval forest approach. This program takes pre-determined variants listed in a data file that contains the nucleotide change and its position and predicts if the variants are deleterious.The output from the SnpEff tool is display are displayed in the table as:
1. SNPS
2. variants impact HIGH
3. variants impact LOW
4. variants impact MODERATE
5. variants impact MODIFIER
6. variants effect 3_prime_UTR
7. variants 5UTR_premature_start_codon_gain
8. variants effect 5_prime_UTR
9. variants effect downstream_gene
10. variants effect initiator_codon
11. variants effect intron
12. variants effect missense
13. variants effect splice_acceptor
14. variants effect splice_donor
15. variants effect splice_region
16. variants effect start_lost
17. variants effect start_retained
18. variants effect stop_gained
19. variants effect stop_lost
20. variants effect stop_retained
21. variants effect_synonymous
22. variants effect upstream_gene
23. Labels
The labels are the classes or the groups the genes are mapped into.The labels can act as both target variable or feature as per the need of the user for solving their specific problem
23.1 No Label
This selection is provided to enable users to view the properties of all genes without labeling them into different gene categories or annotations. This is to let users examine the features of multiple genes and identify common patterns among them. As it involves the inspection of all the genes therefore they work only for "Submit for analysis" button .
23.2 Classical Genes
Classical genes can be defined as the most well studied genes mainly for their visible mutant phenotype (for example: liguleless3).
23.3 Pan-genome Genes
A gene in a given taxonomic group is either present in every individual (core), or absent in at least a single individual (dispensable).
23.4 Origin Genes
Gene duplication is an important evolutionary mechanism allowing new genetic material and thus opportunities to acquire new gene functions for an organism. There are different Origins of duplications such as whole-genome duplications, tandems, etc.
Graph interpretations:
To the top right corner of the plots/graphs, there are options to download plot,
zoom-out/zoom in, reset axes, autoscale, toggle spike lines, show closest data on hover, compare data on hover, box select,pan
and lasso. Users can also select specific legends to view data only for the selected legends. Details on the interactive plot options are available here:
Interactive graph features
1. Histogram
The Histogram shows the frequency distribution of the selected Variant annotation feature such as SNPS, variants impact HIGH, variants impact LOW etc. The X-axis in the histogram represents the range of values present in the selected Variant annotation. The Y-axis represents the frequency of values. In addition to the graph,to increase the interpretability of the data, We have also included P-values,mean and standard deviations of the selected datasets.
2. Count and distribution
The Count and distribution plot is a smoothed, continuous version of a histogram estimated from the data. The most common form of estimation is known as kernel density estimation.The x-axis is the value of the selected Variant annotation just like in a histogram and the y-axis in a distribution plot is the probability density function and not a probability. The difference is the probability density is the probability per unit on the x-axis. In general the y-axis on the distribution plot is a value only for relative comparisons between different categories like classical and other genes.
3. Pair plots
The pairs plot builds on two basic figures, the distributions and the scatter plot. The distributions on the diagonal allows us to see the distribution of a single selected Variant annotation while the scatter plots on the upper and lower triangles show the relationship (or lack thereof) between two variables such as SNPS and variants impact HIGH.
3. Box plots
Boxplots are a measure of how well distributed the data in the selected Variant annotation's are. It divides the data set
into three quartiles. This graph represents the minimum, maximum, median, first quartile and third quartile in the
selected Variant annotation . It is also useful in comparing the distribution of data across data sets by drawing
boxplots for each of them such as the core, non-core, dispensable and private genes. Boxplots can be used to:
- Identify outliers or anomalous data points
- To determine if our data is skewed
- To understand the spread/range of the data
4. Violin plots
Violin plots are similar to box plots, except that they also show the probability density of the selected Variant annotation
at different values. These plots include a marker for the median of the data and a box indicating the
interquartile range, as in the standard box plots. Overlaid on this box plot is a kernel density estimation.
Like box plots, violin plots are used to represent comparison of a variable distribution such as
SNPS, variants impact HIGH
(or sample distribution) across different "categories" (Classical/Other or Core/Non-core) .
A violin plot is more informative than a plain box plot. In fact while a box plot only shows summary statistics
such as mean/median and interquartile ranges, the violin plot shows the full distribution of the selected
Variant annotation .
5. Joint plots
A Joint plot comprises three plots. Out of the three, one plot displays a bivariate graph(scatter plot) which shows how the
one variable( such as SNPS) varies with the another variable(such as variants impact HIGH). Another plot is
placed horizontally at the top of the bivariate graph and it shows the distribution in the form of
marks along an axis for each one of the selected Variant annotation's ( for example rug plot of SNPS and variants impact HIGH).
The third plot is placed on the right margin of the bivariate graph with the orientation set to vertical and
it shows the distribution of again the two selected Variant annotation's .
It is very helpful to have univariate and bivariate plots together in one figure.
This is because the univariate analysis focuses on one variable, it describes, summarizes and shows any
patterns in your data and the bivariate analysis explores the relationship between two variables and also
describes the strength of their relationship.
6. Scatter plots
A Scatter plot is a great way of exploring relationships or patterns in data. But adding a regression line can make those patterns stand out . Therefore Scatter plot with simple linear regression for the selected Variant annotation's, explains the strength of the relationship between the two variables such as SNPS, variants impact HIGH in your scatter-plot using R2, the squared correlation coefficient.It is always between 0 and 1. Higher R2 indicates stronger relationship between the selected Variant annotation's .
7. Correlation plots
Correlation heatmap is graphical representation of correlation matrix representing correlation between different selected Variant annotation's. Correlation ranges from -1 to +1. Values closer to zero means there is no linear trend between the two variables. The closer to 1 the correlation is the more positively correlated they are; that is as one increases so does the other and the closer to 1 the stronger this relationship is. Correlation plots also alerts us to potential multicollinearity problems.
8. Downsampled Dendrogram plots
A dendrogram is a type of tree diagram showing hierarchical clustering — relationships between similar sets of
genes based on the selected Variant annotation's. They are frequently used in biology to show clustering between genes
or samples, but they can represent any type of grouped data.The dendrogram is built on downsampled data to save
time and complexity.
The branches in the dendrogram are arranged according to how similar (or dissimilar) they are. branches that are
close to the same height are similar to each other; branches with different heights are dissimilar —
the greater the difference in height, the more dissimilarity. Also the different clusters of genes based on the
selected Variant annotation's are marked by different colors.
9. Downsampled Hierarchical Scatterplot
The Hierarchical scatter plot is a type of pair plot that can be used to visualize the relationship of different pairs of Variant annotation's on the clusters of genes formed during dendrogram. Therefore to create Hierarchical scatterplot we need to input the number of clusters in the choose cluster input box. For the Downsampled Hierarchical Scatter plot, we need to specify the number of clusters we want to view. Moreover, this number of clusters corresponds to the number of clusters formed in the Dendrogram plot, so we perform this analysis only after studying the Dendrogram plot. The Downsampled Hierarchical Scatter plot is dynamically sized for the number of clusters, so every user can analyze the dataset based on how many clusters they believe the selected data is forming .
10. Downsampled Hierarchical Heatmap
Hierarchical clustering heatmap is an intuitive way to visualize information from complex data. It’s also called
a false colored image, where data values are transformed to color scale. Heat maps allow us to simultaneously
visualize clusters of samples(selected Variant annotation's) in the column and features (genes) in the rows.
First
hierarchical clustering is done of both the rows and the columns of the data matrix. The columns/rows of the
data matrix are re-ordered according to the hierarchical clustering result, putting similar observations close
to each other. The blocks of ‘high’ and ‘low’ values are adjacent in the data matrix. Finally,
a color scheme is applied for the visualization and the data matrix is displayed. Visualizing the data
matrix in this way can help to find the variables(selected Variant annotation's) that appear to be characteristic
for each sample cluster.
10. Heatmap
A heatmap is a plot of rectangular data as a color-encoded matrix.This is a great way to visualize data, because it can show the relation between variables (selected Variant annotation's) including genes. The heatmap here shows the relation between the first 100 genes and the selected Variant annotation's values to reduce time and complexity .
PCA
PCA reduces the high-dimensional interrelated data to low-dimension by linearly transforming the old variable
into a new set of uncorrelated variables called principal component (PC) while retaining the most possible
variation.
The first component has the largest variance followed by the second component and so on.
The first few components retain most of the variation, which is easy to visualize and summarize the features
of original high-dimensional datasets in low-dimensional space. PCA helps to assess which original samples are
similar and different from each other.
In our case when the selected number of Variant annotation's are higher than 2 then, it is arduous to visualize them at
the same time to interpret the genes.PCA transforms them into a new set of variables (PCs) with top PCs
having the highest variation. PCs are ordered which means that the first few PCs (generally first 3 PCs but
can be more) contribute most of the variance present in the original high-dimensional dataset. These top
first 2 or 3 PCs can be plotted easily and summarize the features of all original variables (selected Variant annotation's).
11. PCA 2D variables cluster
A PCA 2D variables cluster plot shows how strongly each characteristic or variable influences a principal component. In the plot we can see these vectors(variables) are pinned at the origin of PCs (PC1 = 0 and PC2 = 0). Their project values on each PC show how much weight they have on that PC.
12. PCA 2D observation cluster
A PCA 2D observation cluster plot showing clusters of samples/observations based on their similarity.PCA does not discard
any samples or observations. Instead, it reduces the overwhelming number of dimensions by
constructing principal components (PCs). PCs describe variation and account for the varied influences
of the original characteristics. Such influences, or loadings, can be traced back from the PCA plot
to find out what produces the differences among clusters.
Observations further out are either outliers or naturally extreme observations. Plot observations are annotated with shapes
and colors to highlight gene models and gene types for example (Classical/Others). In this way, we can easily hover
over any data point, especially outliers, to find gene models with extreme observation values across various sample types.
(selected Variant annotation's)
13. PCA 2D biplot cluster
PCA biplot = PCA observations plot + variable plot.
we can probably notice that a PCA biplot simply merges an usual observations PCA plot with a plot of
variables. The arrangement is like this:
*Bottom axis: PC1 score.
* Left axis: PC2 score.
In other words, the left and bottom axes are of the PCA observations plot — use them to read PCA scores of the
observations (dots).
Another nice thing about these plots: the angles between the vectors/variables tell us how
characteristics or variables correlate with one another.
* When two vectors are close, forming a small angle, the two variables they represent are positively
correlated.
*If they meet each other at 90°, they are not likely to be correlated.
*When they diverge and form a large angle (close to 180°), they are negatively correlated.
14. PCA 3D variables cluster
This is similar to the PCA 2D variables cluster plot but now instead of two principal components, we can have three PC'S that contribute to most of the variance present in the the original high-dimensional dataset. These top 3 PCs can be plotted easily in 3 dimensional space and summarize the features of all original variables (selected Variant annotation's) .
15. PCA 3D observation cluster
This is similar to the PCA 2D observation cluster but now instead of two principal components, we can have three PC'S that contribute to
most of the variance present in the original high-dimensional dataset. These top
3 PCs can be plotted easily in 3 dimensional space and summarize the variance in the observations .
A PCA 3D observation cluster plot displays how much variation each principal component captures from
the data. If the first three PCs are sufficient to describe the essence of the data.
https://pcingola.github.io/SnpEff/