Research Design and Interpretation
The Scientific Method
Urry et. al., Campbell Biology, 12th Edition
The essence of biology and research follows the scientific method —a generalized procedure delineating the steps to take when conducting an experiment. Why? Since the very purpose of an experiment is is to answer a question, test an hypothesis. The scientific method is the foundation of which nearly all good experiments are constructed from, helping to ensure validity from results.
For example, a main way one can determine if their tested is through controls: positive and negative controls. Negative controls are relevant, identical tests that are guaranteed not to work while positive controls are guaranteed to work (thus, making it harder to come by). This is based off the underlying principle that 1) we cannot assume anything when running an experiment and 2) we need to compare our data to determine significance.
How to read a scientific Paper
The copious number of research papers being published each day is a testament to the widespread use of the scientific method. But these papers are often far more convoluted than the simple experiment shown in the figure. This is due to the fact that scientific papers are technical documents designed to convey research findings or developments to another scientist in their field of study, typically not the general population. Consequently, the information is not as digestible, but this shouldn’t underscore the relevancy of this content — these papers often reveal new discoveries and findings of science!
Scientific papers typically start with an abstract, or a general overview of the paper. Along with the title, these components are often the most critical in determining if a paper aligns with your goals. The bulk of the paper contains the main text and data, the methods (the technical aspects allowing the expirment to be reproducible), the discussion (where interpretations and conclusions from data are listed), and references.
Below are the four main types of scientific papers you may encounter:
Research Article: Similar to the aforementioned content, this paper retells original research and their findings.
Review Paper: This type of paper summarizes findings throughout a field of research, but offers no original research.
Methods Paper: A novel protocol or procedure is established.
Clinical Report: Data is determined and examined from clinical trials or case studies.
The x-axis, log-fold change, indicates how differentially expressed genes are and are usually in base 2. For example, Rbp1 is 32 times (2^5) more expressed than the control. Red coloring indicated overexpressed while blue coloring indicates underexpressed. The y-axis is p-value which is a measure of significance.
Data Visualization
Data visualization is arguably one of the most essential parts of research, as figures are not only expensive to add in papers but also critical in enhancing human interpretation of results. As a result, there are countless types of graphs that can be used to represent. Examples include histograms (segregating data into bins then recording counts/frequency as bars), manhattan plots (plotting association values of loci on chromosomes with certain traits), scatterplots and PCAS (useful in identifying clusters), box and whisker and violin plots (illustrating quartiles and data distribution, and volcano plots and heatmaps (frequently used to visualize gene expression — what genes are expressed are expressed and how much). There are also some other funky yet extremely useful methods such as clustering (used to associate cells with similar gene expression patterns) and dimensionality reduction (reducing variables into “UMAPs”).
