Life sciences are some of the most important fields of science today. Throughout the disciplines of biology, biomedicine, and pharmaceutical sciences in pharmaceutical and biotech companies, biostatistics plays a key role. We use it to analyze the data from experiments, improve study designs, interpret the results of studies, and make decisions within all these areas of life science. Biostatistics is applicable in all of these areas, and more, because it allows us to understand the underlying processes that you may be investigating.

While biostatisticians are essential in many areas, from biology and medicine to public health, understanding biostatistics is critical for other professionals in these areas, too.

If you are performing an experiment, conducting a study, or are interested in life science analytics, you will need to analyze the data to make conclusions or get insights from it. Biology and biomedical professionals will encounter biostatistics in most areas of their careers.

When reading almost any life science research publication, you will need to understand how to read biostatistics to understand the results. This is essential for both biologists and biomedical professionals who want to stay current with the latest research statistics for the pharmaceutical industry to discover biomarkers or therapies for patients.

Biostatistics enables us to understand and analyze the data or results we get from experiments, research, or observations. This is one of the reasons why the biostatistical field is important not only for biostatisticians, but also for doctors, biologists, epidemiologists, public health decision-makers, bioinformaticians, health data scientists, and other professionals from most life science branches.

The next subsection will help you understand the specific areas of life science where biostatistics is used. This is very important as every life science area is different and requires a different approach to resolve the research problems

Biostatistics in human life sciences

Biostatistics is essential in many human life sciences. Epidemiologists heavily rely on different types of data to infer their insights. Understanding statistical concepts is essential to understanding population-level biological events and helps both doctors and public health professionals in their work. One such example is the past SARS-CoV-2 pandemic. You must have heard about concepts such as reproductive number (R) or SARS-CoV-2 cumulative incidence of infections, mortality, lethality, excess deaths, and other similar terms. All these concepts are derived using biostatistical concepts and formulas.

Epidemiology is predominantly used in biomedical science areas by public health professionals to make decisions for disease response and to keep the population as safe as possible.

Figure 1.1 – Areas of human life science where biostatistics is used

Medical doctors need biostatistics, not only for their everyday work but also for publishing their academic work, which generally utilizes statistics to summarize and analyze the data of studies and to understand novel discoveries in their profession by interpreting study results from novel publications.

Biomedical research is one of the areas which is heavily reliant on biostatistics and knowing biostatistics is important, not only for statisticians but also for biomedical researchers. Even with access to expert biostatisticians, it is helpful to understand biostatistical thinking and analysis methodology to help with discussions on study design, analysis, and interpretation.

Pharmaceuticals, research, and the development of medications are among the largest industries today. The majority of advanced research in these areas is vital for the biomedical industry.

Biostatistics for biology

While statistics itself is used in many different areas of science, its application in biology has evolved in a specific way due to the nature of different biological domains. Statistics cannot be effectively applied without knowing the basic principles of these biological disciplines.

The following figure shows biostatistics applications in different areas of biology:

Figure 1.2 – Areas of biology where biostatistics is used

Bioinformatics relies on different statistical methods and algorithms combined with computational tools to process and analyze large amounts of biological data, such as RNA (ribonucleic acid) sequencing data (transcriptomics), DNA (deoxyribonucleic acid) data (genomics), and many other data types. Bioinformatics is specifically focused on genetics and molecular biology but implements methods such as biostatistics and machine learning.

Ecological studies are one of the examples where biostatistics is one of the main biological research drivers. Analyzing plant and animal populations, trends, dynamics, and relations between organisms and their environments would not be possible without biostatistics. Next, we will discuss biostatistics applications in different fields in more detail.

Biostatistics in epidemiology and public health

Epidemiologists and public health professionals answer some of the most important public health questions but also make decisions in different communities. They investigate diseases and events in smaller groups of people, cities, and countries, or even worldwide phenomena, such as pandemics. All this would not be possible without the use of biostatistics facilitating the process of analyzing the biomedical and population data.

Epidemiologists often create different statistical models to try to relate infectious outbreaks to causes and then prevent future infections and isolate the infection source. One such example is studying types of food ingested by infected individuals and identifying a potential bacterial or viral food source, or a location, such as a hotel or restaurant, as a source. Biostatistical models are often used in identifying the sources of infectious agents, which will be discussed in more detail in later chapters.

A few biostatistical concepts used in epidemiology and public health are as follows:

• Prevalence
• Cumulative incidence
• Identifying causes for infectious outbreaks
• Characteristics of microorganisms causing outbreaks in a population
• Epidemiological monitoring of populations
• Decision-making based on biostatistics

Biostatistics in medicine and biomedical research

Medicine and biomedical research are very active sciences today, as they directly or indirectly impact almost everyone’s life today. These two disciplines rely heavily on the use of biostatistics. It is of the essence not only for medical doctors but also for biomedical researchers.

Medical doctors’ understanding of the probability of different diseases or outcomes is highly dependent on understanding the statistical concepts and how these apply to groups of patients.

Here are some of the most important concepts used in biomedical research:

• Understanding of incidence of safety for treatments
• Making conclusions about the symptom and disease relations
• Creating biomedical studies
• Analyzing biomedical data
• Interpreting novel research
• Becoming specialized in biomedical data analysis

Biostatics in zoology and botany

A significant portion of the research in biological disciplines, such as zoology and botany, depends on quantifying different aspects of their behavior, life cycles, relations with their environmental factors, and many other aspects.

Some examples of areas in zoology and botany that apply biostatistical methods are as follows:

• Animal behaviors
• Plant growth
• Relations between animals and their environment
• Relations between plants and their environment
• Biochemical composition of different tissues in animals
• Biochemical composition of different tissues in plants
• Identifying feeding patterns in animals

Biostatistics in ecology

Ecology is one of the life science disciplines significantly based on biostatistics. Understanding the population’s diversity and the relationships between organisms, as well as the relationships between organisms and their environments, is facilitated using different biostatistical methods.

Some important areas of the use of biostatistics in ecology are as follows:

• Relationships between animals and their environment
• Relationships between plants and their environment
• Studying biochemical and molecular aspects in zoology and botany
• Studying relations between humans, ecology, and environmental protection

Biostatistics in pharmaceutical research and design

The pharmaceutical industry is one of the main drivers of research and innovation today. Biostatistical analyses enable pharmaceutical companies to design, conduct, and make decisions based on different analyses and insights. In fact, almost any high-quality research project in the pharmaceutical industry consults biostatisticians to make sure that the design is statistically sound and that it can answer the research questions to drive forward the development of assets and to conform with regulatory requirements. Biostatistics is also the key to analyzing adverse events from the data collected during a study, which is essential for any pharmaceutical product. All medications are required to have a list of adverse effects and this is something that can be seen in everyday life. Biostatistical calculation of incidence rates is one of the ways to assess those adverse effects.

Biostatistics is used to assess the efficacy of different therapies and, as such, is a key element in selecting the candidate drugs for diseases such as diabetes or cancer, which are then further evaluated in clinical trials using different biostatistical methods.

Calculating required sample sizes for pharmaceutical studies is a common task of biostatisticians within the pharmaceutical industry, but this is also intertwined with trial design and endpoint selection.

Here is a summary of the uses of biostatistics in pharmaceutical R&D:

• Creating R&D studies
• Evaluating drug safety
• Selecting drug candidates through biostatistical screening
• Designing clinical trials
• Evaluating results
• Research publications
• Meta-analyses of therapy effects
• Regulatory submission

Biostatistics in bioinformatics and genetics

Molecular biology is one of the biological branches that is very specific in terms of using statistical analyses. From structural biology to analyzing gene expression, biostatistics plays one of the most important roles in bioinformatics. Statistical bases form many genetics areas, such as inheritance genetics and population genetics. Here are some of the areas of bioinformatics and genetics where biostatistics plays a pivotal role:

• Differential gene expression
• Structural biology
• Mutation biology
• DNA analytics
• Mendelian inheritance
• Mendelian randomization studies
• Population genetics

To be able to perform statistical analyses in life science and research, you will first need to learn how to address scientific questions in these areas. Scientific questions are a way to define what it is that we are trying to understand or what goal to achieve. In this chapter, you will learn by example how to formulate scientific questions related to various fields related to biostatistics, such as biomedical research, before any relevant statistical analysis is made. One of the first questions to answer is, “What is the goal of a statistical analysis?” This goal is closely related to different life science aspects, therapies, biological processes, or genetic characteristics, and in this section, those will be covered in more detail.

Once scientific questions are made, they are then used to formulate different scientific hypotheses. The main characteristic of any hypothesis is that it can be tested and there is an alternative (opposite) hypothesis to the main one. So, the baseline scenario assumption can be that there is no statistically significant result, and we can test the alternate scenario: that there is a significant result against the baseline or null scenario. We can call the null hypothesis H0 and the alternate hypothesis Ha.

How to formulate scientific questions related to diabetes

The effect of different lifestyles on the outcomes of type 2 diabetes mellitus has been debated for decades.

Let’s pose a couple of scientific questions about diabetes. We will use the letter Q for scientific questions:

• Q1. Is body weight related to type 2 diabetes mellitus?
• Q2. Are there other risk factors for type 2 diabetes mellitus among those studies?
• Q3. Which of the lifestyle factors is the most important risk factor in type 2 diabetes mellitus?

Now, let’s formulate these questions even better. We will mark formulations using the letter F:

• F1. Null hypothesis (H0): Body weight is not related to type 2 diabetes mellitus.

  Alternate hypothesis (Ha): Body weight is related to type 2 diabetes mellitus.

• F2. Null hypothesis (H0): There are no other risk factors for type 2 diabetes mellitus among those studied.

  Alternate hypothesis (Ha): There are other risk factors for type 2 diabetes mellitus among those studied.

• F3. This question will not have a null hypothesis as it is already assumed there are risk factors in the questions. So, the goal of answering this question is to compare the risk factors and identify the most important one. This would be an observational scientific question.

So, why do we usually formulate the null hypothesis as a negation of what’s being tested? Well, we want to know the following: Can I show evidence that contradicts that baseline negative assumption? If I can, then I can reject the null hypothesis. If there isn’t enough evidence to negate the null hypothesis, I can say that I cannot reject the null hypothesis (avoid the mistake of saying that no evidence is evidence of a null hypothesis).

How to formulate scientific questions related to cardiovascular disease

Is ST (the last wave on the electrocardiogram of the heartbeat) elevation closely related to heart disease? With this, we move to the following questions:

• Q4. Do cigarettes increase the risk of cardiovascular diseases?
• Q5. Is an ECG closely related to cardiovascular disease?
• Q6. Are there any other risk factors for cardiovascular disease among the studied parameters?

  Let us make a more structured formulation as follows:

• F4. Null hypothesis (H0): Cigarettes do not increase the risk of cardiovascular diseases.
• Alternate hypothesis (Ha): Cigarettes increase the risk of cardiovascular diseases.
• F5. Null hypothesis (H0): ECG is not closely related to cardiovascular disease.
• Alternate hypothesis (Ha): ECG is not closely related to cardiovascular disease.
• F6. Practice yourself!

How to formulate scientific questions in biology

Here are a few examples for formulating questions in biology:

• Q7. Learn to explore which genes are highly suppressed in lung cancer.
• Q8. How similar are the genomes of mice and humans?
• Q9. What are the differences in plants and minerals collected from localities A and B (Ca, Mg, K)?
• Q10. Does water temperature affect plankton?

Practice formulating these questions as hypotheses or concrete study questions!

You may find the answers at the end of Chapter 1.

It is generally believed that biostatistics is mostly about numbers and graphs. The reality is quite different. Biostatistics is also about understanding life science problems and finding ways to resolve those using statistical methods. There are six main problem-solving skills in biostatistics:

• Helping life science professionals resolve research problems in these domains through the use of data
• Helping life science professionals interpret the results of their research
• Making sure the published research is both statistically and biologically valid
• Helping R&D professionals make decisions in the projects
• Revealing objective truths about different phenomena through the use of data
• Explaining the abstract features of mathematics and biology in an intuitive and easy-to-understand way

One of the most important impacts of biostatistics is transitioning from statistical knowledge to actual problem solutions in life sciences. This will be discussed in more detail in the rest of this chapter.

Biostatistics is needed to derive insights from life science experiments and convert measurements and observations to life science solutions.

Professionals in life science and biostatisticians, working together, design different types of experiments, measurements, and observations. All these can be written or stored as data. Data is a source of information from those experiments, measurements, and observations.

Data can originate from observations, too. One example of observation is the diagnosis by a dermatologist or the identification of species by biologists.

Biostatisticians are there to help make sure this data is valid and make it meaningful. Further, data should be organized and structured, often presented in the form of tables to be prepared for further analysis and interpretation.

To make the data useful, we must understand all the details about the data and how these are related to domains where biostatistics is applied. One of the most important aspects of biostatistics is the context around the data. This context can significantly affect the results and is one of the reasons why biostatisticians are more specialized in life science domains than general statisticians.

One of the main goals of biostatistics is to take all available inputs in the form of data and process them in such a way as to produce meaningful insights, answers, and conclusions and provide information to make decisions in life science.

Here is the biostatistics workflow:

Figure 1.3 – Biostatistics workflow

There are two main types of data: numerical (for example, the measurement of the hemoglobin level in blood in which we are using numerical values such as grams per liter or g/L) and categorical, such as a doctor’s diagnoses of their patients in a form; “Yes” for a positive diagnosis or “No” for a negative diagnosis. These types of data can be further divided into subcategories, which will be discussed in detail in the next chapters.

Understanding data sources is essential for biostatistics. Biostatistics is focused on statistical models but also on domain knowledge and, as such, has evolved as a separate branch of both statistics and life sciences.

This book will provide many different examples that will show you how to use biostatistics specifically for different domains, such as diabetes research, cardiology, and biostatistical studies. Further, in this chapter, we will discuss how the Python programming language can facilitate the implementation of biostatistical methods.

Biostatistics and Python

Most biostatistical analyses today are implemented in some form of software or a programming language. I chose Python as a programming language for this book for several reasons. Python is one of the most advanced languages for data science and biostatistics. As programmers today are moving toward using Python, keep in mind that it is one of the most wanted skills in most areas that have to do with analytics. Libraries such as Biopython and SciPy are among the more than 100,000 libraries that make Python so versatile, meaning that almost any biostatistical analysis can be performed using this programming language. It is open source, meaning it is transparent and free for anyone to use.

The following figure is an example of using Python for biostatistics:

Figure 1.4 – Biostatistics and Python

Its integration with advanced machine learning and bioinformatics algorithms gives a biostatistician a whole new spectrum of approaches and provides the most advanced frameworks for using biostatistical algorithms at this time.

Finally, the most important part – learning Python through a portfolio of practical projects provides you, as a reader, with two important qualities: being able to use one of the most wanted programming languages out there can be beneficial for your career, and having a portfolio of more than 10 practical projects using biostatistics and Python provides significant resources for your portfolio as someone who plans to use or advance your career by using biostatistics.

(A stands for answer)

• A6. Null hypothesis (H0): There are no risk factors for cardiovascular disease among the studied parameters.

  Alternate hypothesis (Ha): Risk factors are present among the studied parameters.

• A7. This question would have no concrete hypothesis. Instead, the overall goal of the study is to identify the genes that are highly expressed in lung cancer tissues..
• A8. We can re-formulate this question into three potential options based on different levels of similarity:
  • The mouse-human genome similarity is low (0-50%).
  • The mouse-human genome similarity is medium (50-90%).
  • The mouse-human genome similarity is high (>90%).
• A9. To answer this research question, we can formulate it as follows:

  Are Ca, Mg, and K concentrations higher in locality A compared to locality B?

• A10. Null hypothesis (H0): Water temperature does not affect plankton.

  Alternate hypothesis (Ha): Water temperature affects plankton organisms.

Keep practicing yourself!

In this chapter, you learned about the needs and uses of biostatistics in life sciences, how to formulate the research questions, and how the Python programming language can help you with that. You also learned about the specific applications of biostatistics in biology, clinical research, and the pharmaceutical industry.

In the next chapter, you will learn in detail how to install and get started using the Python programming language.