Wk07-Midterm Review

This paragraph introduces a midterm review video for the spring of 2024. The speaker outlines the topics to be covered, starting with the properties of a normal distribution. Key points include the symmetry of the distribution, its continuous nature, and the importance of the mean and standard deviation in defining it. The speaker also clarifies misconceptions about the mean and standard deviation of a normal distribution, emphasizing that not all normal distributions have a mean of zero and a standard deviation of one, which is a characteristic of the standard normal distribution. The paragraph concludes with a transition to the next topic, which involves comparing read depth and count-based data with normal distributions.

The speaker discusses the differences between normal distributions and count-based data, such as read depths in sequencing. They explain that read depths are discrete values and positive only, which violates the symmetry of a normal distribution. The speaker also mentions that count-based data is never truly a normal distribution, as it is discrete and often modeled by the negative binomial model, which accounts for the process of sampling finite locations multiple times. The paragraph highlights the complexity of accurately modeling biological data and the move away from normal distributions in advanced models.

This section delves into the evolution of sequencing technologies, contrasting first and second-generation sequencing. The speaker explains that first-generation sequencing was limited to reading single sequences at a time, requiring separate physical solutions for each sequence. In contrast, Next Generation Sequencing (NGS) allows for massively parallel processing at a nano scale, significantly increasing efficiency. The speaker also discusses the challenges of genome assembly, particularly with repetitive sequences, and how long-read sequencing has overcome these issues by providing reads longer than the repetitive regions, enabling more accurate genome assembly.

The speaker emphasizes the importance of visualizing biological measurements using two-dimensional graphics like scatter plots, as opposed to one-dimensional graphics like histograms. The reason is that two-dimensional graphics can reveal interactions and correlations between different measurements, providing insights into the biological system. The speaker also touches on the use of data frames in R and how they are identified, including their structure and properties.

The paragraph discusses operations performed on data frames in R, such as extraction of specific rows and columns. The speaker explains that extractions do not simplify if there are multiple containers, preserving the integrity of the data. They also provide examples of how data frames can be manipulated and the importance of understanding the class of each object in the outputs, such as character vectors, data frames, and numeric vectors.

This section focuses on the order of operations in R, using an example of calculating the average miles per gallon for straight engine cars from the 'mtcars' dataset. The speaker advises thinking through the order of operations in terms of what needs to happen for the next step to occur and suggests using parentheses for clarity. They also explore the concept of logical vectors and how they can be coerced into numeric values for calculations, providing various methods to determine the number of false values in a logical vector.

The speaker presents a task of extracting rows from the 'empty cars' dataset where certain conditions are met, such as the 'vs' column being equal to one and the 'horsepower' being less than 100. They demonstrate the use of logical vectors and the '&' operator to combine conditions. The paragraph also includes an example of plotting data from the 'iris' dataset, focusing on rows where 'petal length' or 'petal width' exceed certain values, and the importance of understanding conditional extractions for data analysis.

This section discusses error handling in R when dealing with logical checks and function assignments. The speaker explains that R produces an error if a logical check receives more than one value, as it requires a single true or false outcome. They also explore the concept of function assignment to objects and the importance of understanding when and how functions are executed within an R session.

The speaker presents a physics problem involving the motion of an object under constant acceleration, which can be solved using R. They demonstrate how to define a function for calculating displacement based on time, initial velocity, and acceleration. The example involves finding the time when an object thrown into the air returns to the starting point, showcasing the application of R for scientific calculations.

This paragraph covers the use of regular expressions (regex) in R for pattern matching and string manipulation. The speaker provides examples of constructing regex patterns to match specific criteria, such as containing an 'e' followed by an 'm', or starting with an 's' and ending with an 'n'. They also explain the use of 'gsub' for replacing parts of a string based on a pattern and the importance of understanding regex symbols for effective text processing.

The speaker discusses advanced techniques for data visualization in R, particularly focusing on the use of 'ggplot2' for creating complex plots. They address common issues such as layering plots and adjusting the order of plot elements, such as ensuring that points appear on top of violin plots. The paragraph also touches on the use of aesthetics to map variables to visual properties, enhancing the clarity and information conveyed by the plots.