PhD applications

https://lucylai.com/blog/gradapps

PhD applications (my sister’s guide)

https://steep-wavelength-0f6.notion.site/Applying-to-Clinical-Counseling-Psych-PhDs-Open-Access-Materials-Tips-Pitfalls-f5fb95046ede406b9baa70b319bbe96d#2ec8d15412db4b4eab10a821825f6525

PhD applications

https://github.com/womanreading7/PhD-Application-Guide

PhD interview tips

https://www.reddit.com/r/gradadmissions/comments/1q3wdjq/from_a_harvard_neuro_phd_how_to_prep_for/

This particular guide will be most applicable to those applying to STEM (particularly molecular biology) programs in the United States of America. These types of programs tend to have rotations and committee-based admissions. As I am from the USA I do not have much insight into PhD applications in Europe/Asia. The guide is quite long, so just scroll to the parts that most interest you. I took the approach of including tons of material so that it may potentially serve as inspiration when you draft your own personalized answers to essays and interview questions.

Thank you to Katelyn Hahn and Tapasvi Bhatt for providing feedback that helped improve this article!

As some background on me, I did my undergraduate degree at Amherst College where I majored in math and biology. I had several short internships in developmental biology, computational biology, and immunology, and wrote a senior thesis in pure math. I spent two years as a research technician at Harvard Medical School/Brigham and Women’s Hospital in Richard Sherwood’s lab. I was applying to genetics, computational biology, bioinformatics, and umbrella biology PhD programs. I ultimately decided to join Stanford’s genetics PhD program. I am just starting my PhD (as of 2026) so any thoughts on “why do a PhD” should be taken with a grain of salt.

At a high level summary, my admissions process in 2025 looked like this:

Applied to 15 programs

Interviewed at 6

Accepted at 4

Should I apply for a PhD?

This is quite a personal decision, and the opportunity cost of doing a PhD is very high. I think if you are 1) in a financial position to do so, 2) are very intellectually curious about a given field, and 3) have good coping mechanisms to deal with the emotionally challenging aspects of a PhD that will all serve you well. Here are a few of my thoughts on the calculus for deciding whether or not to pursue a PhD.

You do not have to pursue a PhD to have a career in science.

For one, there are careers in science policy, investing, healthcare, engineering, consulting, etc. that are available if you have a bachelor’s degree in the sciences. In biotech/biopharma, it is very possible to have a career as a scientist without a PhD. A great example of someone who is a successful life scientist that did not get a PhD is Dean Lee, who writes about breaking into biotech careers without a PhD and is otherwise a phenomenal resource. In my own experience, I have met multiple professional scientists who have advanced high up the leadership ladder in big biotech companies without PhDs, so it is a viable path, and may become more popular as biotech companies evolve and funding situations evolve within academia. On the physics side of things, I have friends who have gone down the PhD route and others who have worked at national laboratories like Lawrence Livermore National Laboratory, so irrespective of STEM discipline there are probably opportunities that exist for you to leverage your STEM knowledge, whether in or out of the lab. Also, something that has surprised my pre-med mentees in the past is the fact that if you are an MD, you can still do research. I have met several MDs now that have research labs doing very interesting work, so doing research as a doctor is definitely a possibility, even if you do not go down the MD-PhD route (which is an incredibly grueling path, I remember a statistic that 1% of MDs are MD-PhDs).

• You may want to consider that having a PhD may make you “overqualified” for certain jobs.

• For whatever reason, people are more likely to leave negative reviews on Yelp than positive ones. Similarly, people on PhD subreddits tend to skew more negative about their thoughts on their PhDs. The ones who have made it and are very happy with their PhDs likely don’t feel as compelled to post. Keep that in mind when you are looking online for advice: online forums may skew negative, and you should talk to real people with PhDs for their thoughts if possible.

Possible careers after a biosciences PhD

As I am in the biosciences I know most about potential career paths after a biosciences PhD. The following careers may also apply for PhDs in other scientific disciplines. When you apply to different schools, you will see that about half of PhD graduates for biosciences go down the postdoc route and half do something else, e.g. go into industry, but outcomes really depend school to school. I’ve observed that postdoc/biotech scientist are the most common paths post-PhD across all schools I’ve interviewed at. Here are career paths that I have personally observed for biosciences PhDs:

• Postdoc

• Professor

• Research scientist in academia (not a postdoc, a more long-term position)

• Core facility scientist (e.g. a sequencing core, imaging core, proteomics core, etc.)

• Lab manager

• Group lead within a biotechnology company

• Scientist (wet lab) within a biotechnology company

• Bioinformatician within a biotechnology company

• Field application scientist

• Startup founder

• C-suite executive at a biotech company

• Management consultant (McKinsey/Bain/BCG/etc.)

• Management consultant at a boutique healthcare consultancy (Clearview/Lumanity/ZS/etc.)

• Management consultant for a government consultancy (Booz Allen, etc.)

• Investment banker

• Equity research analyst

• Partner at a venture capital firm

• Government policymaker

• Patent lawyer

• Technology transfer officer

• Science journalist

• Educator (K-12 teacher, etc.)

• Data scientist/ML engineer

  • Dario Amodei is an example, he has a Biophysics PhD from Princeton

A big caveat is that the topics that are in vogue after the duration of a STEM PhD (4-6 years) may be very different from those that are in vogue now, so your skills may be more or less valuable depending on how the market shifts. For instance, now spatial transcriptomics is a hot topic, but in several years there may be a new hot topic being discussed in industry.

Are you competitive for a PhD?

I would ask other people you know if you are competitive for a PhD application. I wasted a lot of time and money in my senior year of college applying to PhD programs when I really was not a competitive applicant. Being a research technician helped me become way more competitive my second time applying to grad school, and many people (roughly 50%) had also spent some time after college doing research prior to interviewing at graduate programs. I was able to be co-first author on a paper from my research technician position which I’m sure helped considerably, but having authorship on a paper is by no means a requirement for admission into PhD programs. You should at least have strong research experience in one lab that you can talk about in detail.

If you are a US citizen looking to apply to a PhD in the biosciences, my opinion is that a general biology master’s degree is not worth the cost, because you can get research experience (the most important part of your application) as a technician in the same amount of time while getting paid, rather than paying tuition. However, master’s programs in bioinformatics are a different case particularly if you did not study bioinformatics as an undergraduate. The calculus is very different for international students though, since a master’s can be useful for establishing a research community in the USA/eventually getting a visa, etc.

At the end of the day, much ink has been spilled on the topic and I am only one person with his opinions, so just do some thinking about it and ask around because the answer depends on your circumstances and preferences. It is most important to ask yourself whether a PhD makes sense and whether it is truly something you are committed to do.

Application writing

How many programs should I apply to?

• I would apply to 10-15 programs. In the year that I was applying (2025), funding for graduate students was particularly low because of funding cuts in the USA, so I applied to more schools than I might have otherwise applied to. The price of applying to a given program averages out to roughly $100 per school, so factor that in when you are putting your list together.

• I applied to several programs that were hosted at one school, e.g. Harvard Biological and Biomedical Sciences (BBS), and Harvard Systems, Synthetic, and Quantitative Biology (SSQB). As far as I know applying to two programs will not exclude you from interviewing at two programs if you are selected to interview. Anecdotally, I only interviewed at Harvard BBS, but during interview day I met people who were interviewing at both BBS and SSQB.

What factors should I consider when selecting schools?

Research fit

• Definitely have 5-8+ faculty you would want to work with. I generally don’t think it’s worth your time to reach out to professors beforehand in biosciences PhD applications.

  • When I reached out to professors the response was “Thanks for applying! Reach out to me when you get into the program.” It was a waste of my time and theirs.

  • The exception is if there is a very strong research alignment, you have a contact with that professor, or your program is not a rotation program. Molecular biology PhDs have rotations where you try several professor’s labs before you commit for 4+ years, whereas PhDs in ecology/evolutionary biology in the USA require you to apply to a specific professor’s lab. Other disciplines that this might apply to are geology/meterology/environmental science. In fact, for direct admit programs the opposite might be true: you may NOT want to express interest in multiple professors in a department. In the latter case, for programs that require you to work with a specific professor, it is essential to communicate with them.

Location

• Don’t apply to a school in an area that you don’t want to live in. For instance, I have a friend who does not like Baltimore, so he did not apply there for grad school.

• Similarly, somewhere closer to home may be beneficial or harmful to your mental health depending on how you set boundaries with family. There may be pressure for you to visit home more often, which could be a negative, but if you benefit from being able to visit family often then that is a positive. It depends on your individual preferences and situation.

• Would you like to do grad school in a big city or small city? That matters too.

• If you are not sure, it is a good idea to reach out to anyone you know who lives in that area and get their unfiltered opinion on the city. If you are selected to interview at a particular location and are invited to fly out, that is also a good way to assess whether you would like living in that city.

Weather

• If you get seasonal affective disorder (SAD) you should think about whether going to somewhere like Seattle is right for you since it is rainy for much of the year. Similarly, if you are from southern California and do not want to move somewhere it snows, east coast schools may not be good options.

Cost of living/stipend

• Cities like Los Angeles, New York, Boston, and San Francisco have very high cost of living. Make sure the stipend the school offers you will be able to cover your basic needs while in graduate school.

Goals post PhD

• Will the school help you with your goals after getting your PhD? Your goals will almost definitely change, but if you want to be working in biotech after a PhD, it may be quite beneficial to do your PhD in a biotech hub like Boston or San Francisco. Similarly, if you want to end up long-term in a particular location, a PhD may be a good opportunity to set up connections in that city.

Program type/size

• It will probably be easier to get into more specific programs than a general umbrella program. So if you have a specific skill set, like biochemistry or biophysics, apply to those programs because the pool will be smaller. Also, programs have very different class sizes. This can change the vibe of your PhD experience, so just think a bit about whether you want to be in a big class for your PhD or a relatively smaller one.

Classes

• How many classes do you have to take, and how much time will they take up? My recommendation is to look for programs with less classwork, so that you can devote as much time to research as possible.

Boundaries around graduate school

• This really depends on your own preferences, but I think it is best to keep your individual graduate school process to yourself if you are in a workplace where there are multiple people applying to grad school at the same time.

• Try to establish firm boundaries around talking about grad school applications (so not really discussing programs applied to, reading each other’s application materials, discussing resume/GPA, etc.). If not, things can be awkward, especially if a person gets into someone else’s dream program but the person themselves does not, and then people may speculate about why one person got into a given program while the other did not. Also, if your coworkers assume you are competition, that can potentially make the workplace environment hostile.

• To communicate this to coworkers, I would probably use a phrase saying something like “I’d like to keep my grad schools I apply to/interviews confidential as I navigate the application process”.

Tips for essay writing

• Always check your essays to make sure the name of the school is right. Also, make sure that you are spelling the names of the faculty members correctly.

• Avoid making your essays sound like ChatGPT/Claude wrote them (even if they helped you). (https://en.wikipedia.org/wiki/Wikipedia:Signs_of_AI_writing) So remove em dashes, specific LLM coded phrasing, etc as much as you can. I used LLMs to condense word count but not to write large chunks of my essays.

• NOTE: LLMs cannot count words accurately. Do not trust their word counts they provide. Also, note that if you are using Claude, for instance, and you update its memory to avoid LLM phrasing like em dashes, etc., it will likely completely ignore those notes and still produce LLM-sounding text. Definitely thoroughly proofread your essay after running it through any LLMs.

• What I did was I spent a substantial part of my SOP essay talking about my most important research experience, as opposed to spreading out my discussion of different research experiences. I heard this tip from when I was on a panel about PhD admissions at Harvard, and think it is a good general piece of advice.

• Get people to read your essays, it will help if your PI you work with can give it a read, as well as other research mentors you might have.

• Ask for graduate school recommendations at least a month out. You will need three recommenders for many schools. Give the recommenders a spreadsheet of schools so they can keep track of what schools they’ve sent letters to and which ones they haven’t. Follow up if the deadline is coming up and they have not yet submitted the recommendation letters.

• Most schools have a research statement/personal statement, but some schools deviate from that format. Be aware of that fact when you are planning out your essay writing schedule.

For the SOP specifically:

• In the SOP you should demonstrate that you like the day to day of scientific research. If you have experiences that challenged you/stories of adversity, integrate them in your SOP.

• Pick your favorite project to talk about as opposed to scattering your efforts to discuss all of your projects. If you had a few experiences and you want to vignette, make them short and sweet so you get to an experience that you spend at least half if not more of your essay on.

• You could also include a step for what you do next on a project. If you don’t have results yet, just say that! That’s okay. The question is what captures your reader and you, so anchor yourself there! It really pays to pick your favorite story and dig into it.

• Your PI should help you with your SOP. Make sure you give them enough time to make edits and suggestions. If you can get postdocs to read it that is good too.

• For clarity and impact, get people to read it who are outside your immediate area. Don’t write the SOP for specialists, write it for everybody.

• The main advisor’s recommendation letter holds much more weight than the other recommendation letters.

For the personal statement:

• Use the personal statement to show people you work hard, you think about the broader community, and you are generous with your time.

• Don’t repeat the same statements in the personal statement you’ve made in the SOP or your CV.

• Aspirations are really important: e.g. who you want to be as a professional, but what you have done speaks louder about who you’re committed to being than what you might say your dreams are. Try to anchor your aspirations with a lot of detail. If you worked prior to grad school (e.g. a non-scientific career), include it in the personal statement!

Sample Statement of Purpose (SOP)

Here is an example SOP that I wrote for Harvard BBS. I reused bits and pieces of this essay for pretty much every application. In between paragraphs, I’ll explain roughly the function of the paragraph.

Since childhood, I have been drawn to the mystery of how life organizes itself. In high school, while shadowing a developmental biologist, I watched fertilized lamprey eggs transform into larvae, their cell divisions unfolding in a precise choreography. I entered college wanting to address a deeper question: how individual molecules give rise to the organismal whole. Through coursework and research, I learned to observe complex biological systems and perturb them to reveal how their parts interact to sustain their behavior. I aim to build a career uniting molecular and computational biology to understand how genetic variants, gene networks, and protein structures generate the order underlying human biology, and how their disruption leads to disease. The Biological and Biomedical Sciences (BBS) PhD program at Harvard, with its tight-knit graduate student community and rigorous, interdisciplinary training, is an ideal environment to pursue these areas of research.

This paragraph provides a motivation for why I got interested in biology research in the first place. You could use some sort of childhood anecdote for this as well, but it’s important to avoid sounding cliched by tying it into an overarching narrative within your essay. For instance, if you say that you were inspired by a movie or a book you saw as a kid, it is important to then say what that pushed you to do in high school or college, and how that has led you to pursuing a graduate degree. I provide a statement of what sort of research I would like to do in the future, which relates to my previous research experiences, and explain how Harvard fits my needs very briefly.

As an undergraduate, I sought to connect my studies in math, computer science, and biology. I explored developmental biology in Dr. Katerina Ragkousi’s lab and computational biology as a summer researcher at the Cold Spring Harbor Laboratory in Dr. Peter Koo’s lab, experiences that inspired me to combine computational and biological skillsets in translational research. Thus, I joined Dr. Daniel Dwyer’s lab at the Brigham and Women’s Hospital for a summer internship. The Dwyer lab studies allergic response in the distal lung epithelium, a tissue type critical for gas exchange. The cellular mediators of allergy in the lung are poorly understood, and I hypothesized that single-cell RNA-sequencing (scRNA-seq) could illuminate the cell types responding to tissue inflammation. To test this hypothesis, I analyzed scRNA-seq data to identify alveolar type 2 cells as primary responders in allergic inflammation and pinpointed genes most altered in expression, later validating their protein-level changes using flow cytometry. These insights highlighted AT2s as potential therapeutic targets and suggested that their inflammation-responsive genes could serve as disease biomarkers. This experience confirmed my fascination with combining quantitative methods and cell biology and deepened my belief that such integration can supercharge biological discovery.

My experiences from undergraduate research were in very different fields of study and somewhat scattered. Thus, I spent relatively less time talking about them than my technician position ultimately. I wanted to provide a transition from my earlier undergraduate research experiences to my later research technician position, so I discussed how I performed interdisciplinary studies, and got basic exposure to computational biology during summer research opportunities. I spent more time discussing a computational immunology summer internship because it allowed me to perform both wet and dry lab research simultaneously.

For every research experience I mention, I try to provide a basic motivation of why the research matters (a rationale for trying to answer that research question), the methods that I used to address that question, the ways in which I disseminated that research (e.g. a poster or paper), and what I learned from that experience. Finally, I use that sentence to transition to another research experience. E.g. for this computational immunology experience, I mention how since I was working on transcriptomics, I wanted to study it from a mathematical lens, which led me to my undergraduate thesis.

Wanting to strengthen my understanding of transcriptomic data analysis, I pursued a senior mathematics thesis with Dr. Ryan Alvarado reviewing the field of optimal transport, a framework recently used to trace cellular differentiation trajectories from sequencing data. This project deepened my understanding of how mathematical tools can quantitatively measure cell fate transitions. While reviewing the field of lineage tracing, I came across technologies that record developmental history by writing barcodes into the genome. I was thus inspired to pursue genome engineering research, which would allow me to observe and perturb biological systems.

In my current position as a research technician in Dr. Richard Sherwood’s lab at Brigham and Women’s Hospital, I am leading a project investigating how low density lipoprotein receptor (LDLR) mutations affect low density lipoprotein (LDL) uptake. LDLR variants are responsible for most familial hypercholesterolemia (FH) cases, but FH is severely underdiagnosed because of our inability to identify pathogenic variants. To address this problem, our lab conducted a CRISPR screen using prime editing, a versatile genome engineering technology capable of introducing any single-nucleotide substitution. Using this platform, we systematically evaluated thousands of LDLR variants and quantified their effects on cholesterol uptake. By modeling LDLR variant effects, we improve FH diagnosis and pave the way for earlier FH treatment, offering a key case of how genetics can improve health.

Again, I used a transition sentence to move from my discussion of my undergraduate thesis to my research technician experience. I do not assume that the reader knows what LDLR or FH is, and provide information on why studying this gene matters. In the abstract, they would assume that there is a disease called familial hypercholesterolemia, and that variants in the gene LDLR contribute to disease risk, so we should better characterize such variants to address the severity of this disease. I did not spend time explaining the background of the disease very deeply to save on space, because I wanted to highlight my own research contribution.

I explain the core screening platform used to address this question, which uses prime editing, that I briefly discuss. Prime editing is complicated, but I kept the explanation to a bare minimum: all the audience really needs to know is that it can introduce any single-nucleotide substitution, and nothing about the components of a prime editor, the limitations of prime editing, etc. I also provided a sense of the scale of the screen, by stating we quantified the effects of thousands of different variants.

More surprisingly, I also identified novel gain-of-function (GOF) variants from our screen that strongly increased LDL uptake and could protect against heart disease. To validate the top-scoring GOF variants, I independently designed and executed follow-up flow cytometry assays, finding that these variants increased LDL uptake levels to twice that of wild-type LDLR. I also demonstrated that combining multiple GOF alleles has a synergistic effect on cholesterol uptake. I even found that co-installing a defined GOF variant alongside known pathogenic LDLR variants could rescue LDL uptake. Given that a wide variety of LDLR variants can cause elevated cholesterol levels, a defined GOF variant installed concurrently with a LOF variant could serve as a “one-size-fits-all” treatment for high cholesterol.

To identify the mechanisms by which GOF variants increase LDL uptake, I simulated point mutation effect on LDLR structure using computational chemistry tools to generate hypotheses for how key genetic variants change interatom interactions. Utilizing these approaches, I discovered that several top GOF variants potentially generate novel interatom interactions between LDLR and its ligand apolipoprotein B, suggesting a mechanism by which GOF variants may increase LDLR-LDL binding affinity. [a few sentences on future directions, omitted]

Here, I talk a bit about my own contributions to the study, and the implications of my findings. I highlight that I both made contributions on the wet and dry lab front. It is important to provide insights into your thinking as a scientist and how you followed up on your insights if you have the room. While I don’t spend that much time using phrasing like “I wondered” or “I hypothesized”, that is a good way to frame your contributions to a project. In this case, there is a logical flow of how I executed the experiments:

We did an unbiased screen on thousands of LDLR variants

> a few gain-of-function variants are discovered, which is surprising

> genetic screen results can be sometimes unreliable/artifactual, so let’s validate them with an orthogonal assay

> The results validate, can we combine GOF variants together for epistatic effects?

> Can GOFs ameliorate the negative effects of LOF variants if both occur simultaneously?

> Ok, so GOF variants have potent effects on increasing LDL uptake, but how do they actually drive LDL uptake?

> computational chemistry to provide hypothetical GOF mechanisms

Across my undergraduate coursework and post-baccalaureate research, I have solidified my commitment to leveraging genetics, genome engineering, and computational biology to improve our mechanistic understanding of genetic disease. Looking ahead, ​some of the most enticing multidisciplinary genetics problems lie in complex conditions like heart disease, schizophrenia and cancer. Their human toll cannot be overstated, nor can the challenge of treating them. The BBS program’s collaborative structure and diversity of research areas make it a uniquely rich environment for interdisciplinary training. I am particularly inspired by Jonathan Strecker’s work utilizing protein engineering to expand our genome editing repertoire, and Kamila Naxerova’s research using evolutionary and spatial genomics to study tumorigenesis. I believe that my future as a researcher lies in the study of how genetic variation shapes biological systems and how to engineer tools that fine-tune human genomes to be robust against future pathological processes. Pursuing a PhD in Harvard’s BBS program will allow me to bridge molecular and computational approaches and prepare me to lead interdisciplinary research at the frontier of human health.

First, I start my paragraph making a case for why I am a good candidate to be doing mechanistic analysis of genetic disease. I have done coursework and previous research projects in which I explicitly analyzed disease mechanisms using genetic tools. I speculate on future issues the field of functional genomics will face and explain why the BBS program is a particularly good place for me to grow as a scientist. I provide a sentence talking about two professors I would like to work with.

You always want to talk about why you want to go to a specific school or program in your essay

List 1-3 professors of interest in the essay if you can. I do this here. Generally though, don’t go too heavy in an application saying you want to work with a specific person. That’s almost a red flag.

Sample Personal Statement

I went with a different angle for my personal statement for Harvard BBS to showcase my identity as an artist. In every paragraph I tried connecting my creative practice with how I engage with science and the scientific community. I was also able to reuse bits and pieces, if not the whole essay, for multiple applications.

Alongside biology, art has been a steady presence throughout my life. From nervously presenting a tiger painting in an elementary school competition to entering a campus arts exhibition in college, creating visual art has always grounded me. It helps me de-stress after long days of pipetting and has shaped how I approach science. Art has taught me to experiment boldly, learn from others, and consider the human impact of every work, whether on paper or under a microscope.

One lesson I have learned from art is the value of experimenting with media. Over the years, I have worked with ink, graphite, paint, and charcoal, each suited to a different creative vision. Likewise, at different points in my scientific experience, I have tried being a wet lab developmental biologist, a computational biologist, a mathematician studying biology, and now a hybrid wet and dry lab geneticist. Just as an artist chooses the most effective media, I aim to select the right approach for each biological question. My diverse experience allows me to view biological problems from different angles and pursue the most exhaustive path to discovery.

Another lesson I have learned from my lifelong dedication to art is to continually seek learning. To strengthen my art skills after college, I balance weekdays in the lab with weekend classes at Boston’s Academy of Realist Arts studying classical drawing. As I learned how to measure the human figure with a plumb line, I simultaneously sought opportunities to understand CRISPR screening from my coworkers. Both settings reward patience and curiosity, reminding me that every day in the lab and studio is a chance to grow.

Finally, being an artist has led me to think deeply about the impact of my work. Countless afternoons spent studying Impressionist paintings at the Museum of Fine Arts have shown me that technique matters, but so does intent. Similarly, the best scientists I’ve worked with are deeply creative and intentional with their science, understanding that their work can impact people’s lives. I strive to emulate that same ethos in my work. I also value the individual human connections that art can facilitate, as I love to gift friends my art. Science, much like art, can reach people on a personal basis in the same way, not just as a large-scale published project. The same pride I felt holding up that tiger painting returns in the small moments such as teaching a mentee how to hold a pipette to the bigger ones, like when they watch their first experiment work.

I hope to embody the traits of a passionate artist in my work as a PhD student in Harvard’s Biological and Biomedical Sciences program. I want to stay attuned to how my work can benefit human health while continuing to expand my knowledge and technical repertoire in graduate school. And all the while, I would be excited to continue exploring Boston’s unique art scene, while working on my own artistic practice when I am not busy in the lab.

Timeline

Here is a general list of dates that you can keep in mind when writing the applications. Not every successful applicant adheres to such a timeline of course. I think working on the applications in advance if you can will help reduce stress massively later on as you move closer to December.

September: Start writing essays, complete 3-5 applications (no need to send in yet)

October: Request letters of recommendation, complete 3-5 applications (no need to send in yet)

November: Complete 3-5 applications (no need to send in yet)

Thanksgiving weekend: final stretch, crank out 2-3 applications and/or proofread previous applications. Send in most applications. By this point you should have most of your applications ready to submit. DO NOT WRITE ALL YOUR APPLICATIONS DURING THANKSGIVING WEEKEND!!

Early December: Applications due

My personal application process

I applied to the following programs:

Harvard Biological and Biomedical Sciences (BBS)
Harvard Systems, Synthetic, and Quantitative Biology (SSQB)

MIT Computational and Systems Biology (CSB)

MIT Biology

University of Washington Genome Sciences

University of Washington Molecular and Cellular Biology (MCB)

Princeton Quantitative and Computational Biology (QCB)

Tri-Institutional Program in Computational Biology and Medicine (Weill Cornell)

Stanford Biophysics and Genetics

UCSF Tetrad

UCLA Bioinformatics

UPenn Genomics and Computational Biology

Northwestern Driskill Graduate Program in Life Sciences

UCSD Biomedicine

That’s 15 programs in total! I will note that as I wrote some of the essays, I let them sit for a month or two, and as I improved the “skeleton” that I used for the SOP/research statement for different schools, I updated my older essays accordingly.

Interviews

These can be tough, especially if you have a lot of them. I interviewed 6 times, but certainly some people have even more interviews, closer to 10. At this point, there is likely a 25%-80% chance you will get in if you are invited to an interview. This totally depends on the program and year though. For reference, I interviewed at Northwestern Driskill, Harvard BBS, Princeton QCB, UCSF Tetrad, Stanford Genetics, and UCLA Bioinformatics.

Pre-interview tips

• If you have a lot of interviews, check the dates for when they happen, and book travel as soon as possible. Communicate early and often with your professors/advisors to let them know that you will have to miss school/work to interview.

  • Consider whether it is physically possible to attend all the interviews. If not, you will probably have to choose between schools to interview at.

  • I recommend making a spreadsheet to keep track of all of the dates.

  • If possible, take some time off before/after interviews whenever possible to give yourself mental space to destress during the interview process.

    • Try not to go to work on the same day as a virtual interview if you can help it, to ease stress.

    • If you are doing wet lab work and can have your co-workers cover for you while you are away, ask early. E.g. if you have cells in culture that you need someone to take care of, ask them (while giving enough advanced notice), or if you have molecular cloning tasks.

• Glance at the grad school handbook for the programs you are interviewing at. It may be helpful to familiarize yourself with how the grad program functions.

  • If you really want to, you can come up with some classes in the program that interest you, though I was never asked about classes that interest me during interviews.

• Come in with a strong sense of why you are applying to a particular program at a given institution, so you can answer that question confidently if asked. I did not do this for UCSF Tetrad, which was a mistake. (E.g. Why would you apply to UCSF Tetrad vs. BMI)

• If the school releases a list of professors you are interviewing with, read up on them as much as you can. I like doing that because it makes me feel more comfortable and confident going into the interviews.

• Get 7-9 hours of sleep before your interview if you can.

  • Don’t get wasted the night before the interview.

• See if there is anyone you know that is a grad student at the school/program you are applying to and if you have time to get a meal with them.

• I recommend prioritizing which programs you care the most about, and putting the most time into preparing for those interviews. Some schools will tell you the faculty you are interviewing with in advance, some won’t. If the schools do tell you who you are interviewing with, look them up and research them as much as you can so you will have an extra confidence boost going into the interview.

• Check the weather if you are flying out somewhere and dress accordingly!

  • If you are flying, make sure everything you’re bringing passes TSA.

• Bring an extra change of clothes just in case.

• It can be helpful to do mock interviews, but also do not do too many at once, or within a short span of time.

• Definitely have questions prepared for the interviewers. These can vary depending on who you are talking to. If you are interviewing with someone you’d really like to rotate with, try to get a sense of their personality and whether they’d be open to have you rotate

  • Prepare enough that you can have an entire conversation that is exclusively you asking the professor questions (for 30-45 minutes)

• If on a virtual interview, make sure that you have a professional background (nothing you wouldn’t want an interviewer to see), and that your camera angle is good. MAKE SURE you have access to a space that is 1) private and 2) has decent WiFi. If that is not the case, you can look into coworking spaces or libraries.

  • I would test your call with a friend on Zoom/Teams/whatever platform the interview will take place on if possible to make sure everything is in order.

  • If you have a hotspot feature on your phone, test that and if that works, that is a backup option for you in case your service cuts out.

• If you are being flown out somewhere to interview, you will almost definitely have a roommate.

  • Consider bringing earplugs if your roommate snores.

• Don’t let anything PhD students in the program say psych you out before the interview. When I was having dinner with PhD students the day before one of my interviews I was told I might have to give a chalk talk, which didn’t end up happening. By the time you reach the day before the interview, there’s not that much more you can do to prep.

During the interview day tips

• The most important part of the interview day/s are the interviews themselves, not necessarily the auxiliary talks and keynotes. If you have to skip out at any point during the day, skip out during those talks, and get some water/use the bathroom/take a walk/take a power nap on a couch, etc.

• Interviews are less focused on behavioral style questions as in a typical job interview and are instead focused on learning about you as a scientist.

• Interview formats can differ a lot in length and what you are asked. There is usually not a fixed format for the interview unless interview time is really limited.

• Don’t be alarmed if your interview schedule changes because a certain professor cannot make it. This can totally happen.

• Going into my interviews, I was very scared that the interviewers would be more adversarial. Practically every interview I had was not very adversarial, thankfully.

• You will definitely interview with people who are not in your field.

• No particularly technical questions were asked during any of the interviews. What I mean is that I was not asked to draw out any protein structures or solve equations. This may differ based on discipline though. If you are doing hardcore comp bio/EECS you may get some technical questions.

• Don’t stress out if you’re a bit late to interviews. It can be hard to navigate campuses and the interviewers understand. You can sometimes use Google/Apple Maps to find your way to interview buildings if you do not already have a buddy that helps you find your interviews.

• Don’t go to the interviews in a hoodie and sweatpants. Most people will be wearing business casual clothing. No need for formal wear, but if you want to dress formally, you totally can.

  • Thus, there isn’t a need to buy a new wardrobe for graduate school interviews.

• The interviewers likely do not have that much sway over your decision individually. Even if one really liked you that isn’t a sign you are getting in per se. If one of the interviewers is on the admissions committee that will matter a lot more.

• If you are sick during an interview, bring cough drops/drink cough syrup the entire day, and drink a ton of water.

• I would bring two water bottles around the whole day, and a watch.

  • During one of my interview days I was drinking so much water that I drank my whole bottle’s worth, and there were no opportunities to fill it up during the day.

  • The watch allows you to gauge time elapsed while not making it look like you’re bored or not focused on the conversation. This allows you to stay on track with timing for interviews, even though the professors are supposed to keep an eye on the time.

After the interview tips

• Great job on finishing the interviews!! Write down your thoughts on how the interviews and day went immediately after the interviews are done when you have time. Listen to your gut and write down if you felt really strongly about particular interactions, or if something stuck out to you about the program, positive or negative.

• Don’t get wasted after the interview. If you are tired you really do not have to go to any afterparties. There is no pressure.

• Thank the interviewers after the interviews are over via email.

Questions I prepared for

• Tell me about yourself.

• Tell me about your research.

• What was your specific role in your research?

• What techniques or skills are you strongest in, and what do you want to learn next?

• How do you decide what experiment to do next when results are ambiguous?

• Why do you want to pursue a PhD?

• What do you want to get out of a PhD?

• Why X school?

• What do you want to achieve in your career?

• What do you think is the hardest part of research? What is your favorite part of research?

• What’s something you’re worried about?

• What would you do if you had an issue with your PI?

• Tell me about a paper you read recently.

• Why choose you?

• Tell me about a project you are proud of and how you would communicate it to a layperson.

• What direction do you think the field should go in?

• Did you do an undergraduate thesis? If so, tell me about it.

• How do you embrace diversity? Name a time when diversity played a direct part in a situation you had to deal with.

• What might you not get at this program that you want?

• How do you handle negative feedback?

• Weaknesses/describe a challenge or setback and how you overcame it.

• What would your dissertation proposal be called?

• How do you deal with stress/failures/conflict? What do you do for fun?

• Tell me about the time you had a conflict professionally, or a work related ethical dilemma you’ve had (e.g. a difficult moment), or had a problem and worked through it

• If you had unlimited money/1 million dollars, what project would you do?

• What do you want to do with your career? Where do you see yourself in 5 years?

• What other programs are you interviewing at?

• What are you looking for in a mentor?

Something I did not prepare for but is theoretically possible is if a professor asks you to give a chalk talk on your research. This is not common at all but I have heard of this being done in interviews.

Questions I prepared to ask

Questions for graduate students

• How do you maintain work life balance?

• How is your time spent day to day and week to week?

• Do you feel that the PI supports your work life balance?

• What is the mentorship style of the PI?

• How often do you meet?

• Do you have lab meetings/journal clubs?

• How often do you collaborate with other groups?

• What are concrete things the PI has done to support your professional development?

• Why did you choose this program? What would you tell yourself coming into this program?

• What was a highlight during your time in the program so far?

  I personally think it’s a trap to ask questions like “What’s it like to live in this area, how much is the stipend, what is your rent”, etc. A smarter use of your time is to ask the other questions listed above. When they accept you, you can ask all the questions you want about rent, area, etc.

Questions for faculty

• Where do you see your work going in the next 5 years?

• ​​What are the biggest open questions your lab is most excited about right now?

• What qualities do you see in successful graduate students of this program?

I pulled the following questions from this excellent Reddit thread which I used when I was interviewing.

1. If I joined, what project would you actually want a first-year student to start on?

2. What does a strong first year look like in your lab, in concrete terms?

3. How do you mentor day-to-day, and how often do you meet 1:1 with trainees?

4. How do you help trainees choose and refine a thesis direction once they rotate or join?

5. How do you think about project ownership, collaboration, and authorship expectations?

6. What does your lab culture look like in practice, especially around feedback, independence, and work hours?

7. What resources here make your lab’s work uniquely possible, such as cores, clinical access, datasets, animal models?

8. What skills do students most often need to build when they join, and how do they build them successfully here?

9. Where do your trainees tend to go after graduating, and how do you support those paths?

Questions for the head of the department

• What direction do you want to take the department in?

• What goals do you have for the department?

Questions my research tech PI asked me (mock interview)

I asked my PI when I was a research tech to mock interview me. Here are some of the questions he asked.

• What did you do in your undergraduate thesis?

• What is optimal transport (the topic of my undergraduate thesis)?

• Why did you choose to do a math thesis?

• What are the benefits of using optimal transport over other methods?

• Can you give an example of how you used computational tools to address a biological question?

• Why do you want to do a PhD at Harvard BBS? What fields are you interested in?

• Do you think there will be a point where we can just plug in a variant into a computer and see what it does computationally, so there’d be no point to doing wet lab experiments?

• How close are we to having genome editing treatments in the clinic? Where do you envision this field moving?

• Can you speak more to the clinical classification aspects of the project?

Questions I was actually asked during interviews

This is not a totally comprehensive list. I compiled it based on the questions I remember being asked from the interviewers immediately after the interviews. Clearly, some of these are specific to the project/s and ideas I discussed in my interview, but hopefully they serve as inspiration for what kind of questions you might be asked yourself. I’ve bolded the questions that I think are particularly crucial to have an answer for.

Motivation & Fit

• Why do you want to do a PhD?

  • Why do you want to do a PhD in a specific field?

• What would you want to do in a PhD?

• What do you want to get from a PhD?

• What do you want to do after a PhD? / in 5 years?

• Why X school?

• Why did you choose to do math and biology in college?

• Why did you choose to go to Amherst College?

• Whose labs are you interested in for a PhD?

• What skills/areas would you like to expand into during your PhD?

Research background

• Tell me about your research/previous research experiences

• Why join Rich’s lab (my research tech PI)?

• What did you work on in Rich’s lab?

• Why join a translational lab as a tech?

• Why UCSF Tetrad as opposed to UCSF BMS?

• What are the therapeutic applications of the immunology research you did as an intern?

• Could you talk to me briefly about your two publications?

Technical & Computational Background

• What is your computational/biology background? How do you envision that looking in your PhD?

• Any examples of how you used computation in your work?

• What are the state of the art programs in rare variant analysis?

• What data are rare variant analysis models trained on?

• Do you think, if we tried to use our most updated list of machine learning based tools, that we could predict gain-of-function variants well? Why or why not?

• What did you do in your math thesis?

• How do you think your mathematical background has prepared you for a bioinformatics PhD?

• Where do you see math fitting into your research? Does it generally frame your thinking or does it directly play into your research?

LDLR prime editing project

Note: this was the project I talked about the most in my SOP and during my interviews. Some of the professors had a very good understanding of the project (e.g. they were geneticists, did CRISPR screens, etc.) and some were not familiar with the techniques used.

• What was the most challenging part of the LDLR project?

• What were the variants and what were they doing?

• What did the GOF variants do?

• Did you try to see the effect of combining GOF variants together?

• Did you do testing to determine that there was no off-target editing happening anywhere in the whole genome of the cells?

• How high is your prime editing efficiency? Is that high relative to what the field can achieve?

• What is the readout of the screen?

• What are follow-up experiments you could do for that project?

• What are the clinical implications of your LDLR PE screen?

• What are the implications of your GOF PE screening? Are they biological, technological, or therapeutic?

• How would you try to study the evolutionary variations in LDLR? Why have GOF variants in LDLR not appeared? Wouldn’t they be favorable evolutionarily?

• Why did you use the cell type you used for the CRISPR screen?

• Would you deliver the LDLR therapy to all cell types in the body?

• Would your LDLR therapy work in parallel with PCSK9 inhibitors?

• What is the economic benefit of having an LDLR gene therapy based on GOFs? It seems like statins and PCSK9 inhibitors work well enough already.

• What are the genetics of familial hypercholesterolemia? Is it caused by family history or is it mostly sporadic?

Scientific Interests & Vision

• What are some questions in biology that you find interesting?

• Are you more driven by developing technologies or by a specific field of interest, say metabolism or neurodegeneration?

• What kind of technologies would you want to develop in a PhD?

• Where do you think the genome editing field is going in the future?

• What kind of balance would you have between wet and dry lab work?

• Do you want to do protein design in your PhD?

• How did you find out about the protein design seminars you attended for fun?

Reflective

• What’s a project you enjoyed? What was your contribution?

• What’s a challenge you had in research?

• What was a scientific result that surprised you?

• What are you most proud of in life?

• What does being a scientist mean to you?

• What do you think the difference between collaboration and competition is?

• What makes you prepared for a PhD?

• What kinds of communities have you been a part of in the past?

• What kinds of teaching/mentoring opportunities have you been a part of?

Deciding on a school

Congratulations! Assuming that you have multiple programs to choose between, how do you make the decision?

I considered the following factors:

• Well-being

  • Do grad students seem happy?

  • How many hours do grad students work a week?

  • Do grad students have friends and hobbies?

  • Is it easy to make friends/find communities?

  • What is the general culture of the program?

• Research fit

  • How many professors could I see myself working with?

  • Are people willing to collaborate?

• Resources

  • What kind of equipment exists for scientific research at the institution?

  • What are the computational resources like (e.g. computational clusters, GPUs)

  • What resources exist for mental health services, career counseling, networking, etc.

• Location/Weather

  • How close is this location to home?

  • How do I travel around in this location?

  • How is the dating scene in this location?

  • What is the weather like?

  • What sort of activities/landmarks are in the vicinity of the school?

  • Is this location close to major cities?

• Cost of living/stipend

• Goals post PhD

  • Are there resources to support career development?

  • Is the alumni network strong?

• Program type/size

  • How many people are expected to be in my class?

• Classes

  • How many classes do I have to take?

• TAing

  • Do I have to TA?

I highly recommend reaching out to people from the programs you were admitted to and asking about some of these topics. I think among this list, gauging research fit and well-being are paramount. Those should be some of the most decisive factors for your ultimate decision for where you do your PhD.

Red flags for programs

How do you determine whether a program is NOT right for you? I would look at the list of factors to evaluate a program above and think about whether the program strongly does not align with your interests/values in particular aspects.

• Talk to the current PhD students! The more actively current PhD students are engaged in recruitment events during in-person/virtual PhD interview events, the better. They may say things about their work life balance, or maybe about the department.

  • Are they genuinely happy?

    • Green flag: They like their research, lab mates, cohort mates, community, the city, have hobbies, etc.

    • Red flag: “I think the biochemistry community is weak at X school”

    • Red flag: “I feel isolated and that I don’t have anyone I can talk to about my PhD challenges”

  • Are they complimenting the program but qualifying their statements often?

    • Red flag: “I like this aspect of the program, buuut…”

  • How is their work-life balance?

    • Red flag: “I’m working X hours per week and I’m not happy about that”

    • Red flag: “I have to work weekends often/can’t take time off”

• During the interviews with faculty, you should gauge their willingness to mentor you and help you grow as a scientist.

  • If you are deciding between programs, you can have a follow-up conversation with the faculty members you are interested in working with/rotating with after interviewing!

  • You can potentially reach out to their lab members (particularly former lab members) to gauge if their lab is a good place for you to be in long-term.

My personal rationale for picking my PhD program

Here is a rationale for why I decided to go to Stanford.

• I am from California and thought it would be good to be closer to family.

• I really like the research topics that are being explored at Stanford, particularly at the intersection of technology development/machine learning.

• After talking to several current students, I got the impression that PhD students on aggregate are happy, have healthy work-life balance, and have time to explore their hobbies regularly.

• Stanford has a lot of PIs who are interested in doing a hybrid of wet and dry lab biology. Not too many programs are like that.

• Stanford labs tend to be on the smaller side.

• Stanford has considerable compute resources.

• Stanford is affiliated with the Arc Institute.

• The interaction with biotech and the bigger world outside of academia is unique. There are a lot of opportunities to explore the business side of things or the startup world.

• Stanford Genetics PhD students do not have to TA.

Thank you very much for reading this article and I hope this is helpful to you as you decide whether to pursue a PhD or to navigate your PhD application process!

Label maker is absolutely essential. Having some sort of database for plasmids, oligonucleotides, glycerol stocks is essential.

Having some sort of system for the lab to replenish plates, media, etc. in TC is really helpful.

Having make or break experiments is much more helpful, and doing small experiments leading up to a larger one is really important: in case the small experiments fail, you don’t have to do the long, expensive, time consuming experiment (fail fast). I’ve heard this advice from several successful scientists.

Derisk your PhD by pursuing multiple projects: Pursue a safer, more straightforward project, get it to publication, while spending maybe 20% of your time on a more high risk high reward project.

I think I’ve gone away from the idea that math will directly help biology in the traditional systems biology sense. I spent my undergraduate degree searching for ways to connect mathematics and biology, and I don’t think the connection is that direct. Maybe we are not far along enough biologically to really meld math and biology at a high level, but even my thesis topic (optimal transport), which I thought was really well chosen, was completely useless for my research technician years. That being said, signal processing could be quite helpful for certain biological applications, though more on the electrophysiology or imaging side of things, not necessarily genetics/genomics per se.

There is a classic historic trend of us assuming things are going to be simpler than they are in biology: Once we discovered the genetic code (Watson/Crick/Franklin), we assumed that we would be able to fully decode the language of life. Once we discovered how to sequence genomes, we thought that would allow us to cure all genetic diseases. Once we discovered how to edit genomes, we imagined that we could make designer babies. All of these futures have not come to pass. Similarly, in the context of systems biology, we believed we could use simple mathematical models to study biological systems, which can be done elegantly in certain contexts like in genetic circuits (e.g. Michael Elowitz), but are in all likelihood not particularly useful for studying more complex biological phenomena. Although books like On Growth and Form really captured my imagination as an undergraduate, perhaps looking for mathematical beauty in nature was a naïve way to understand biology.

My opinion is that math can be useful insofar as it can be used to motivate better machine learning models for studying biology. Currently the dominant paradigm for progress in biology is initiatives to generate massive amounts of high-quality biological data and machine learning to tease out patterns from data using massive amounts of compute. Biology is inherently bottlenecked by the time of maintaining living organisms or conducting experiments and will probably be harder to make advances on than pure comp sci/math research. Novel computational methods can help biology, but if the experimental data is not collected it is impossible that computation will help. At the moment, I’m really inspired by folks like Gabe Rocklin, as well as folks I’d want to work with at Stanford like David Cox or Polly Fordyce. Their research focus aligns with my goal of large-scale biological data generation.

(To be fair, it must be said that blindly generating data for training models or training models on the wrong sort of data will also be a waste of time. Refer to (https://www.nature.com/articles/s41592-026-03120-y) as an example: seems that single-cell foundation models do not obey scaling laws. However, seems like ESM models and Profluent models for protein design do obey scaling laws. Particularly from Stanford, professors have perhaps been overselling the capabilities of foundation models like Evo/Evo 2, but Kenny Loi was able to use Evo 2 to do some interesting analysis for his VIPR paper from Jennifer Doudna’s lab.) https://alignbio.org/blog/do-scaling-laws-hold-for-protein-language-models/

One of the things my former PI Rich emphasizes: Make sure our screens are providing value above and beyond what is possible computationally.

As we went through the LDLR PE project, our experimental screening is as good as AlphaMissense it seems. Those models are going to improve faster than the rate at which we can do prime editing screening. The idea to do DMS to find pathogenic variants is not that helpful as is. Obviously if there are compelling cases to do DMS when computational models can’t perform well, that’s compelling. The GOF variant angle is also compelling.

How we personally saw that our screens were providing value beyond what is possible:
Finding GOF variants, finding variants whose effects are mediated through RNA splicing

Cloning

Always make master mixes for PCRs if possible. Use Q5 Master Mix if possible.

Always pipette your PCR, restriction digest, or Gibson up and down 5+ times with a pipette set to ~50% of the reaction volume to ensure that the reaction has been mixed properly prior to incubation in a thermal cycler.

Make sure you are using the right antibiotic for your use case (either the antibiotic for bacteria like ampicillin, kanamycin, etc. or selection agent for cells, e.g. hygromycin/geneticin/puromycin).

Always transport restriction enzymes, PCR master mixes, and Gibson master mixes in -20 coolers.

If using plasmid DNA, 5-10 ng of DNA input is sufficient for a PCR. You can use 100 ng of DNA as input into PCRs from genomic DNA, and 250 ng of cDNA as input into PCRs.

Make sure your restriction enzymes are stocked in advance. You can look for single or double cutters when designing a cloning strategy.

Always check your cloning strategy in silico with Snapgene or related apps.

Always use the NEBuilder calculator if possible for calculating volumes for Gibson assembly. You can scale down if you would like. Scaling to 10 uL or 5 uL should be completely fine.

For the first time doing a transformation when cloning some insert into a vector, always transform the digested vector to test for background. Optimally you will have 10x more colonies on the desired construct plate than the negative control (digested vector).

If you need to concentrate your DNA a bit more, during QIAGEN gel extraction you can elute in 20 uL of EB instead of 30 uL of EB. It works fine.

Make 1000 uL aliquots of dH2O for cloning.

Make glycerol stocks whenever you can. Label exactly where the glycerol stocks are stored in -80. You can make 50% v/v glycerol in H2O and use 250 uL of that with 750 uL of turbid bacteria for making stocks.

In my experience, NEBuilders for 15 minutes work totally fine for 2-way and 3-way NEBuilders, as NEB advertises.
If you are not worried about the amount of colonies you need, you can cut the outgrowth step short during the transformation (e.g. to 30 minutes instead of 1 hour).

Make single-use aliquots of NEBuilder Master Mix. Freeze-thawing it a lot is not advisable.

Always make 5% of the total restriction digest volume the restriction enzyme. For restriction digests, you can design a strategy using the following link:

https://nebcloner.neb.com/#!/redigest.

For ascertaining how long certain restriction enzymes will work for, check this link: https://www.neb.com/en-us/tools-and-resources/usage-guidelines/restriction-endonucleases-survival-in-a-reaction

In general, if missense or indels occur in irrelevant regions of the plasmid, you can proceed with using that plasmid for downstream applications.

Sometimes, if you want to make short DNA amplicons, you can use 3 or 4 oligos and assemble them together using Q5 master mix, without an external DNA template supplied.

If you want to make a custom sequence, you may have to order a block from Twist. We tended to resuspend them at 10 ng/uL in TE buffer.

For choosing primers, if using plasmid DNA as the template, there is a lot of flexibility with the primers to choose, but they should generally be roughly 20 nucleotides or so long. If using gDNA as the template, choose primers with Primer3 using the protocol from Rich’s lab.

For a typical PCR reaction, 50 uL for 25 cycles should be sufficient. If a lot of DNA is required, one can perform a 100 uL reaction.

To extract a desired fragment of DNA, gel purification, PCR purification, or SPRI bead purification may be used. If there is significant primer dimer, use gel purification. If that is not a concern, use PCR purification.

NEBStable cells are preferred for transformation with plasmids for lentiviral transduction because they suppress recombination events. Accordingly, they must be grown at 30 C instead of 37 C. https://blog.addgene.org/preventing-viral-plasmid-recombination

For midipreps, one can pick one colony from a plate and drop it in 50 mL of LB+antibiotic (e.g. Ampicillin). After overnight incubation, the high-yield protocol for QIAGEN midiprep can be used to extract plasmid DNA.

For maxipreps, it is preferred if you inoculate a preculture (3 mL of LB+antibiotic) and then pour into a larger culture (e.g. 250 mL culture of LB+antibiotic).

Library cloning can be extra finicky. Make sure you are using the high efficiency transformation protocol if using NEB cells. If DNA quality is an issue, gel purify AND PCR purify the plasmid. Sequence verify the plasmid using whole plasmid sequencing. If necessary, SPRI bead purification can be used on the PCR-amplified library. The water or primers may be an issue, so make new water aliquots and order new primers if necessary. The primer sequences used may be off, so check that they are exactly as you intended them to be on the order sheet. A one nucleotide difference can make a huge difference, surprisingly enough.

Pre-labeling tubes can help you save time. When doing a lot of minipreps, I would prelabel a set of cryovials (glycerol stocks), 2 mL tubes (miniprep steps), and 1.5 mL Eppendorf tubes (elution), and add 250 uL of 50% v/v glycerol to each one of the cryovials.

When using glycerol stocks, keep them on dry ice if possible and move quickly to ensure they do not thaw. Scratch the top with a pipette tip (preferably P10 tip) and eject into desired LB+amp media.

When trying to triage large amounts of colonies at once, one can pick a colony using a P10 tip, swirl it in a PCR tube containing a full PCR reaction with desired primers and then a preprepared PCR tube filled with 250 uL LB+antibiotic. Repeat this process for all colonies to evaluate. Run colony PCR using the PCR reactions and leave the preprepared PCR tubes with LB+antibiotic in 4C. Check which colonies to proceed with if the correct amplicon size is observed in colony PCR, and inoculate 3 mL tubes for miniprep from the 250 uL LB+antibiotic. A multichannel pipette can be used during the colony PCR and inoculation prep steps.

If you don’t want to start your cultures shaking that have been inoculated yet, you can stick them in 4C until you want them to.

Tissue culture

Always use midi or maxi DNA as input for transfections/transductions.

It can be nice to use your own devoted set of tips to avoid contamination.

Always make sure there are thawed aliquots of FBS you can use to make media + FBS.

Never pass your gloves over an open container. Position all objects in the hood near the back of it, while taking care not to obstruct the airflow.

Take a quick look at the other people’s dishes around you if contamination becomes an issue.

I like freezing down 3 x 10^6 or 5 x 10^6 cells per vial.

Selection with puromycin takes 3+ splits, blasticidin takes 3+ splits, and hygromycin takes 4+ splits.

HepG2s double ~every 2 days

K562s double every day

HCT116 cells double every day

HEK293T cells double every day

THP1 cells double every day

When running cells for flow cytometry, always bring compensation controls.
E.g. If you want to work with cells that are blue+red, bring cells that are blue+red, cells that are blue, cells that are red, and cells that are not fluorescent.

Compensation is not necessary if you are working with RGB fluorophores.

To check positivity for a particular color, I just set an interval gate based on the negative cells, and then check to see what % of the cells to study are to the right of the interval gate.

Genome editing

Major genome editing modalities

What are the major genome editing modalities now? Here is my general view:

Cas9-mediated genome editing

Gene knockout via indels

Specific mutagenesis/introduction of indels

DNA insertion via homology-directed repair (HDR)

Base editing

ABE (A>I>G) adenosine converted into inosine via deaminase (TadA-derived), inosine converted into guanosine

CBE (C>T)

Base editing outcomes predictable by BE-HIVE

Prime editing

Insertion of specific sequences (upper limit of 30-40 nucleotides or so insertion?)

Prime editing pegRNA efficiencies predictable by PRIDICT/PRIDICT2.0, which supports HEK293 (MMR-deficient) or K562 cells (MMR-proficient)

Large genomic insertions

PASSIGE (David Liu)

DAISY (Le Cong)

INSTALL (Ben Kleinstiver)

Bridge recombinases (Patrick Hsu)

Cell killing
Cas12a2

Epigenetic editing

CRISPRi

CRISPRa

Click editing (Ben Kleinstiver)

Tips on genome editing

You can measure Cas9-mediated editing via TIDE/ICE.

You can measure base editing via EditR.

How do people make better genome editors?

Directed evolution (PACE?)

Appending rationally chosen peptides/proteins to current genome editors (e.g. La protein, PE7)

Protein engineering via ML (ProteinMPNN, PE8)

How do people predict prime editing outcomes?

PRIDICT2.0/ePRIDICT (Gerald Schwank’s lab)

OptiPrime (David Liu’s lab)

High throughput screening/functional genomics

What is the right modality to answer your biological question?
Cas9 screening? Base editing? Prime editing? CRISPRi? CRISPRa? Installation of different sequences in a safe site?

Look for the outliers (points that have particularly high or low functional scores) and see if they have a biological explanation for why they have such scores. Also, note the actual magnitudes of the scores. How significant are they? Are they mildly significant, or very significant? How many significant hits are there?

Generally, do trends make sense? Do the functional scores line up with pathogenicity measures (e.g. AlphaMissense, ESM-1v) or evolutionary scores (PopEVE, PFAM)?
Do scores of interest make sense in light of any structures for proteins being studied (if you are performing a DMS on the coding sequence of some protein)? Try drawing a color gradient using PyMOL/ChimeraX.

Always have some positive and negative controls in a high throughput screen.

Positive controls are variants to install that are expected to facilitate a strong effect, e.g. mutating a cysteine residue in LDLR in a LDL uptake screen. Negative controls are not expected to facilitate a strong effect, e.g. a non-targeting sgRNA in a CRISPR screen or a pegRNA that installs a synonymous variant.

It is helpful to have a pegRNA that installs two different variants simultaneously: One installs the desired variant for a DMS and the other variant is a silent mutation that prevents re-nicking by the nCas9 on a prime editor (also, allowing sequencing to distinguish between unedited and edited reads, as there is a low error rate inherent to Illumina sequencing)

Display technologies

Possible display technologies include:

mRNA display

cDNA display

Click display

Ribosome display

PICASSO

CasPlay

Phage display

Yeast display

Mammalian cell display

Bead display

Tsuboyama et al. 2023 successfully used cDNA display to study protein stability in high throughput format.

In vitro protein synthesis kits are way less efficient at synthesizing protein than bacteria.

For screening proteins in high throughput format, I have seen Ben Kleinstiver’s lab use lysate from HEK293T cells.

Cardiovascular disease

HDL-C is cholesterol on an HDL particle. LDL-C is cholesterol on an LDL particle. LDL and HDL are large assemblages of protein and lipids.

For studying protein affinities, surface plasmon resonance (SPR) and biolayer interferometry (BLI) are generally accepted techniques.

The field largely does not use “good cholesterol” or reverse cholesterol transport as a way to clear cholesterol from arteries. CSL-112 failed in clinical trials to produce significant differences in cardiovascular events between placebo and treated, but APOA1 treatments could make a difference if administered for a longer period of time. Rich thinks APOA1 mRNA therapy could make sense. Note that PCSK9i and statins both function through upregulating LDLR. Anyone with familial hypercholesterolemia (FH) will be helped by these treatment modalities up to a certain point (particularly for homozygous FH).

APOA1 Milano failed in clinical trials (potentially Stage 2).

Could we find APOE GOFs? That could be useful for preventing Alzheimer’s disease.

Generally speaking, it seems that variants in genes like PCSK9 might be more helpful for lowering heart disease risk than PCSK9i later in life. The earlier you start treatments for lowering cholesterol, the better protected people are.

The field of APOA1/HDL molecular dynamics is relatively mature. Much less so for LDL

Cholesterol efflux assays are generally done with J774 macrophages or THP-1 derived macrophages. They use fluorescent cholesterol (e.g. BODIPY-cholesterol) or radiolabeled cholesterol (H3-cholesterol).

Vibe coding

Vibe coding is going to be huge. I think at the point that we joined it was relatively new and the term had just come into popular parlance. There need to be guidelines about vibe coding that are established for casual use in science: e.g. no putting original data or novel findings into LLMs.

Also, best practices around using LLMs for things like coding would be good to establish. I have talked to numerous PhD students who are 100% dry lab and they say they have not written a single line of code in 6+ months. If this is the case I see no reason why wet lab biologists must spend significant amounts of time writing code from scratch. At the same time, it is particularly risky for wet lab biologists to do so since they may not come from a coding background.

I suggest prompting LLMs and then asking them to explain the code line by line if you do not understand the code and establishing sanity checks in the code throughout to make sure it is functioning correctly. There is a biostatistician at Harvard and a computational biologist that anyone in the community can book for ad hoc questions. I would suggest future techs and students do so if questions arise. If any lab members can benefit from taking Harvard’s comp bio courses it’s a good idea to pay for them (e.g. single cell RNA-seq data analysis)