#I’m so busy with my thesis/dissertation or whatever and I can’t divide my focus

[bellesans-merci]

“Sakura,” Kakashi called to his sweetly humming wife, pointing with utter bemusement to the two boys huddled against each other on the couch with faces so flushed they were nearly steaming from their ears. “Why are Naruto and Sasuke sharing a shirt?”

The woman deftly slicing vegetables at the counter didn’t so much as spare him a glance, though a pleased smile pulled at the corner of her lips as she deposited carrots and onions into a bowl.

“They were fighting,” she said simply, a shoulder lifting so carelessly he almost felt dumb for asking, “and we do not fight in this house.”

His lips twitched as he read the stretched out words printed boldly on the red cotton, wondering who she’d sweet talked into making an extra-extra-extra large shirt that said “I LOVE YOU” and just how she’d managed to strong-arm the two idiots into it.

He bit down on his lip to keep the laugh bubbling in his chest from spilling over, until the door to their guest bathroom opened and Sai calmly walked into the living room dwarfed by a pink shirt with, “I’m sorry ☹️” embroidered in white across the chest.

The boy looked so unbothered and yet, was very obviously burning at the tips of his ears, that Kakashi couldn’t contain the loud bright laughter that startled his team and beaded tears at the corners of his eyes.

He laughed, and he laughed.

He laughed as he took his place at the stove, he laughed as he donned his puppy printed apron, he laughed at the playful nudge of Sakura’s elbow in his side.

Kakashi’s heart thumped with so much joy and laughter, that the rush of it was almost electric. It warmed him from the apples of his cheeks to the tips of his toes, sparking in his chest like the eye of a storm, and it felt a bit like breaking just as much as it swelled in completion.

This, this was a home—this was his home.

He swiped his fingers at his eye and told himself that the moisture there was from the humor found in the betrayed glares of three pre-pubescent boys, and not because he’d been chasing this feeling all his life.

[cancerbiophd]

Hi Julia! I'm just starting my PhD, and my supervisor gives me a lot of freedom and independence. While I appreciate that, I'm not sure where to start tackling my project. I've done a lot of reading, but I'm having trouble designing good experiments. Do you have any suggestions?

Hi anon!

I have a post here outlining one possible path to take when designing your thesis project–however, it does involve communicating with the PI almost every step of the way. I do think that PI involvement is pretty crucial for the development of a PhD student, so if there’s a way for you to have weekly one-on-one meetings with your PI, I think you’ll find that would be really beneficial. 

In the meantime, let’s talk about experimental design. It’s a very important skill we learn during grad school, and one of those skills that separates someone with a PhD and someone without. It’s part of the “Ph” part of the degree, after all! So if you’re not getting the mentoring necessary to learn how to properly design experiments, then that’s a huge foundation of your PhD that’s missing. If your program (or any related programs) has a class on experimental design, it would be worth taking (I took two different experimental design courses, on top of regular guidance from my PI, and years later I’m still learning so much about the nuances of experimental design). Regardless, as a minimum requirement for a PhD mentor, your PI should be teaching you proper experimental design and results interpretation. Otherwise they’re just a warm money-pumping-lab-having body for the next 4+ years and that’s not what you, a PhD student, are here for, or deserve.

However, I do understand the reality that is busy PIs and large labs. If your PI is really hard to get a hold of, you can try finding other mentors to help guide you, such as: other senior members in the lab (like the lab manager, research specialists, post-docs, even other grad students), your committee members, the PI next door even. I get advice from as many people as I can, because sometimes even if my PI is available, she may not have the best expertise in certain situations. 

As a supplement, I would also recommend finding online resources on experimental design. A quick google search of “experimental design in biology” lead me to this awesome video from Khan Academy that covers experimental design at its simplest. 

Now here’s a quick and dirty 4-step crash-course on experimental design (from my experience in doing biology research):

1) Start with a testable and feasible research question:

This is based off a hypothesis/prediction you make, which in turn is based off the knowledge gaps in your research area

It can be as simple (one experiment) or as complex (aka the focus of your entire dissertation) as you want it to be

It should be testable: you actually have a way to figure out the answer

And also feasible: given your time, ethics, and resources (eg. equipment available, funding, people who can help you)

This is something that reading the literature, or talking to your PI, can help with. 

(Divide up your research question into sub-questions if necessary)

“Yes” and “No” research questions are totally ok. Sometimes it’s as simple as “does my cell line constitutively express this receptor, yes or no?” or “does Treatment X induce my cell to secrete Protein Z, yes or no?”

2) Come up with a method to answer the question:

I like to first go into “fantasy” mode. Like, what would the perfect assay be to answer this? I pretend it’s the year 3050 and whatever I think of we can do. For example: “ah if only i had xray vision and the ability to tell apart a human tumor cell from mouse bone marrow and can see just how many of these tumor cells end up in the bone with my naked eye!” Thinking like this first lets you get to the bottom of “what do we need to solve this problem?” 

Ok, time to go back to the present. We can’t see tumor cells through bone with the naked eye, but what do we have that allows us to do so indirectly? How can I tell the difference between a human cell and a mouse cell, and also quantify that? 

Another part of the design process where reading the literature and/or talking to your PI and other researchers would help with, especially if you don’t know what you don’t know. 

Like if I have no idea that something like intracellular fluorescent labeling and flow cytometry existed to solve my question at hand, I couldn’t even use that in my experiment

Determining the method you will use is sometimes the most time-consuming part imo. If it’s something you or the lab haven’t done before, you’ll need to do a lot of research into the methods (what’s the best reagent? concentration of reagent? do we have access to equipment necessary? do i need specific controls? what’s the specificity and sensitivity of this assay? are there background issues i have to contend with (eg. autofluorescence))? It may take a few tries with optimization before getting the method down for your purposes. And as you can see, it can be super involved, so getting advice and help from your PI or another expert would be really helpful (and time saving!)

3) Design the experiment on paper with the proper variables, controls, and replicates:

I like to pretend I’m solving a murder mystery and I have to convince the jury that Suspect A, with weapon C, is the one who dun it. How do I go about designing an experiment that will eliminate all possible suspects and murder weapons (and thus convince the jury)? 

Sometimes it helps to draw a predicted results graph of your experiment; seeing it in its “final” form may help you realize some controls or treatment combos may be missing. 

Once you’ve designed the experiment, go over it with your PI (or another expert), to make sure it’s sound.

The number of replicates (technical vs biological) you may need will depend on statistical analyses, like power analyses and what kind of statistical experimental design (eg. one-tailed vs two-tailed) you’ve decided beforehand. If this all sounds new, then that’s something you’ll need expert advice on (like from your PI), or take a class in (like biostats), or do lots and lots of independent research (perhaps the most time-consuming and mistake-prone choice). At a minimum though, we always need at least 3 values to perform any stats (so if you’re just running something up the flagpole, n=3 is a quick and dirty thing to do first). 

4) Predict the outcomes of each of your variables and controls and do some thought exercises

Ask yourself if these predicted results will answer your question

If it only answers part of your question, what else do you need?

If it doesn’t really answer your question, what should be changed? 

What if the opposite of what you predicted happens? What would that mean? 

If all this seems super overwhelming, then I think it’s a sign to seek out specific help on experimental design, like your PI or a class. Again, it’s part of your PhD training, but it’s not something you need to, or should, learn all by your lonesome self. 

Good luck with your training and research! I hope you find a good path forward.