8 Tricks to Enhance Your Analysis Reproducibility

Analysis reproducibility ensures that scientific outcomes are legitimate and dependable. Reproducibility is crucial so that you and different scientists can construct upon present analysis and advance scientific data. It additionally builds belief within the scientific technique. Correct labeling, controls, and publishing detailed methodology sections and destructive knowledge make your work extra reproducible and maximize its influence. Avoiding expired reagents, assigning work appropriately, and being aware of exterior components akin to temper and fatigue additionally assist.

You’ve simply accomplished an experiment and obtained your outcomes. Congratulations! You’re one step nearer to increasing the data base and bringing us to a better approximation of the reality.

However how are you going to make certain that your findings are appropriate?

Performing reproducible analysis is vital in science, and there are steps you possibly can take at every stage of your analysis to assist be certain that it’s reproducible. This text discusses eight steps you possibly can take to enhance your analysis reproducibility.

What Is Analysis Reproducibility?

Analysis reproducibility, repeatability, or different comparable phrases, relying on whom you ask, is the power on your findings to be obtained once more by your self and different researchers.

Why Does Reproducibility Matter?

Science is a toolbox for looking for goal fact, and a toolbox is rather more helpful if the instruments are dependable and the particular person utilizing them is constant and is aware of what they’re doing. Analysis is equally extra useful whether it is reproducible, and for a number of causes:

1. Lab mates come and go, and new folks will proceed your work.
2. Different scientists will possible increase upon your work.
3. Irreproducible findings forestall or hinder our pursuit of fact.
4. Reproducible outcomes additional our quest for fact.
5. Reproducible findings construct belief in our toolbox.
6. Trying to duplicate findings identifies damaged or outdated instruments in our toolbox.

How Do I Carry out Reproducible Analysis?

At every stage of your analysis, from starting to finish, there are steps you possibly can take to assist make sure the reproducibility of your findings.

1. Correctly Label All the pieces

Earlier than beginning your experiment, it’s essential to make sure that all of your reagents, supplies, and instruments are accurately labeled. What constitutes correct labeling will fluctuate relying on what’s being labeled, however important data  sometimes added to labels consists of:

• Full substance title, chemical, reagent, buffer, and so forth.
• Key properties akin to focus, dilution, weight, quantity, and so forth.
• Dates on which the substance was acquired and first opened
• Title and/or initials of the one that ready or acquired it

In case your laboratory supplies are improperly labeled, your analysis could also be doomed to be irreproducible earlier than it even begins since you can’t supply or make equivalent reagents and options. To return to the toolbox analogy, think about you’ve been informed to complete screwing in flathead screws. When you unintentionally use a Phillips head screwdriver as a result of it was labeled as a flathead, you probably received’t reproduce the standard of the work accomplished with the correct device.

It’s a lot simpler to inform aside a flathead and Phillips head screwdriver than to inform aside two scientific reagents that look equivalent but differ in unobservable methods. If you wish to set the stage for reproducible analysis, correctly label all the things.

2. Keep away from Expired Reagents or Carry out High quality Management Exams

You’ve ensured you’re utilizing the suitable instruments for the job, however are you aware how lengthy these instruments final? Do the chemical compounds in your lab degrade over time? Have you ever checked whether or not expired reagents change the outcomes you get?

Changing expired reagents with unexpired ones will massively enhance your analysis reproducibility. Producers typically make use of rigorous testing on their provides or reagents. These exams inform expiration dates and shouldn’t be ignored totally. Expiration dates are additionally knowledgeable by (amongst different issues) regulatory necessities and will not completely mirror the reagent’s stability or usability.

Shopping for new supplies additionally prices cash—cash that might not be there if a grant has but to be funded.

Whether or not for that reason or others, you could wind up in a scenario the place you might be utilizing expired reagents.

The excellent news is you possibly can guarantee larger analysis reproducibility when you carry out high quality management exams to verify your supplies’ continued stability or usability.

These exams fluctuate in design, and what they examine. Nonetheless, displaying that unexpired and expired reagents generate the identical outcomes when all the things else is stored fixed can forestall confounding components from impairing the reproducibility of your work.

When performing these high quality management exams, there’ll possible be instances when someone in your laboratory discovers {that a} reagent or substance works completely previous its expiration date. It’s necessary to contemplate that you could be want further high quality management exams as expired reagents proceed to age additional.

High quality management exams themselves take time and sources. Ideally, keep away from utilizing expired reagents altogether. However you probably have no different choice, carry out high quality management exams.

3. Design Your Experimental Protocols in Element

You may have the suitable instruments, they usually’re all working correctly. Now it’s time to design your experiments. When writing your protocols, pay immense consideration to element.

Researchers discover a number of causes to keep away from together with an acceptable quantity of element. I’m certain you’ve heard among the following:

1. “I’ve a extremely good reminiscence, and I’ll bear in mind what I did or learn how to do the experiment.”
2. “We have to reduce down on the strategies part to satisfy the phrase restrict.”
3. “This step doesn’t should be accomplished precisely the identical every time.”
4. “Sure particulars are pointless.”

We regularly give ourselves and our reminiscence an excessive amount of credit score. It’s straightforward to neglect easy steps or combine up numbers in our heads. It’s additionally not the case that you can be the one one to carry out a specific experiment. A cornerstone of analysis is the power of different researchers to duplicate our findings. And different researchers can’t depend on our reminiscence alone.

Concision vs. Element

Now, there are examples of instances when concise writing is necessary, and, at the very least of their present type, journal article submissions are amongst these examples. Take into account although, that concision typically opposes element.

That’s to not say that, regardless of submitting a extra concise methodology as a part of an article, you possibly can’t additionally preserve a extra detailed and exact protocol elsewhere. Good examples of locations it can save you a extra detailed protocol are:

1. Your laboratory pocket book.
2. A devoted binder within the lab.
3. A community drive.

These extra detailed protocols can then be referred to by different members of your lab, and even despatched to different researchers ought to they attain out about an article you’ve revealed. When writing these detailed protocols, it’s necessary to keep away from leaving out the knowledge you’re taking without any consideration and deem pointless.

Right here’s an instance of data that’s straightforward to take without any consideration.

A protocol could state, “carry out the experiment utilizing ten million cells, counted utilizing a hemacytometer”.

This will appear to provide loads of element, because the required variety of cells and the method used to depend the cells are listed.

However have been there any clumped cells if you counted? Did you depend each dwelling and lifeless cells, or solely dwelling cells? How a lot of your cell suspension did you load into the hemacytometer? Did you dilute your cell suspension earlier than counting? What calculations did you carry out to transform out of your hemacytometer counts to your ultimate cell depend?

It’s crucial that particulars akin to these are communicated and shared. Steps akin to these are sometimes assumed to be carried out equivalently between researchers. Particularly researchers working inside the similar lab. Or they’re wrongly assumed to not influence the outcomes.

Nonetheless, it could possibly be the case that utilizing 9 or 11 million cells, attributable to totally different counting strategies, ends in totally different outcomes. Even when steps and particulars appear trivial, it’s higher to be protected and design your experimental protocols intimately than to danger irreproducible analysis.

4. Embody Acceptable Controls and Practices

A part of designing your experiment is selecting which controls to incorporate. With out the suitable constructive and destructive controls, it’s rather more troublesome to find out whether or not the outcomes you observe are genuinely attributable to your experimental variable or confounding components.

Which controls are acceptable on your experiment will fluctuate, however it is extremely uncommon for it to be the case that no acceptable controls exist. Some necessary forms of controls embrace (however should not restricted to):

1. Constructive Controls. Identified to supply the anticipated impact.
2. Destructive Controls. Identified not to supply the anticipated impact.
3. Procedural Controls. Akin to taking measurements twice to scale back potential error.

Along with controls, a number of practices can (and when doable, ought to) be included into experiments to restrict error, forestall bias, and improve reproducibility. These embrace (however should not restricted to):

1. Randomization of group assignments or different components when acceptable. [1]
2. Blinding, or not offering sure details about samples, contributors, or knowledge both to researchers or contributors to forestall biased assortment or interpretation of outcomes. [2]
3. Consistency in in any other case missed experimental parts akin to time of day. Even one thing akin to the intercourse of the researcher may influence animal research outcomes.

Affirmation biases, confounding variables, and different issues that may in any other case trigger your analysis to be irreproducible are rather more simply recognized when you embrace acceptable controls and practices in your experimental design and protocols.

5. Assign Work Appropriately

You’ve written your protocol, you have got the suitable controls, and it’s time to carry out the experiment. Who in your lab will carry out every step? Do you have got collaborators in different labs that can be performing components of the experiment?

Collaboration in a lab is commonly a gateway to success and may deliver collectively totally different experience and ability units to deal with a analysis query. Nonetheless, you will need to allocate experimental obligations appropriately.

In any lab, people can have totally different strengths and weaknesses, and varied ranges of expertise. That is very true in case your lab consists of trainees akin to college students and fellows (at any stage). Whereas it could appear apparent, lab members must be totally educated to deal with any experimental steps that they’re assigned.

For instance, think about you have got a lab member who’s an professional at performing Western blots, and one other lab member who’s an professional at performing PCR. This supplies a wonderful alternative for cross-training and sharing data. Nonetheless, till cross-training is full and each lab members are totally outfitted to carry out each methods, Western blot, and PCR experiments are extra appropriately dealt with by the respective specialists within the lab.

Who’s totally outfitted to carry out totally different methods will change over time as cross-training happens, and lab members come and go. Because of this, it is important to maintain this step in thoughts and assign work appropriately for each experiment.

6. Publish Trustworthy Methodology Sections

Now that you just’ve accomplished your experiments, you’re able to publish them in a journal. Congratulations! As I’ve stated, you could have to shorten your protocols and strategies to satisfy the journal’s phrase restrict and assist the paper circulate or learn higher.

You’ve in all probability learn many papers the place the authors will say one of many following:

1. “In keeping with producer suggestions.”
2. “As per Smith et al.”
3. “As beforehand described elsewhere.”

These methods reduce down the size of the article to satisfy phrase limits and take away particulars which can be maybe pointless. Nonetheless, it additionally introduces the potential to depart out vital methodological particulars or misrepresent the work carried out.

When you intend to explain strategies as being accomplished “in keeping with producer suggestions”, make certain that the producer’s suggestions have been adopted precisely.

Usually, producers present a variety of suggestions and encourage researchers to start inside that vary and optimize as wanted. Different instances, producers will present suggestions for one context, however not the precise context during which your experiment was carried out. These are each examples of situations the place it’s not possible to provide all the main points by describing strategies on this method.

Related issues can come up when describing strategies as being accomplished in keeping with one other article, whether or not your personal or one other. The cited article could not present sufficient element to duplicate the strategies, or it could equally cite an earlier article from which the strategies have been taken, resulting in a reproducibility rabbit gap.

This isn’t to say that credit score shouldn’t be given when strategies are taken from one other research, however the place doable, these strategies must be totally and precisely described. If it isn’t acceptable to incorporate this data in the primary publication, together with it in supplemental supplies or publicly reporting it elsewhere must be prioritized.

As a final resort, if publicly reporting the complete methodology will not be doable, researchers ought to be capable of attain out to you to acquire the complete methodology (bear in mind Tip #3!). Nonetheless you deal with the ultimate publication, you should definitely publish trustworthy methodology sections in order that your work may help us attain a better approximation of the reality.

7. Embody Destructive Information and Appropriate Publication Errors

Researchers typically solely publish constructive findings with out publishing related destructive findings. Publishing destructive knowledge may help different researchers not solely know what to do to duplicate your outcomes but additionally what not to do.

Because the destructive knowledge might not be the sexiest a part of your outcomes, it could make extra sense to incorporate them in supplemental supplies—however they need to nonetheless be included.

After publishing your analysis, together with each your constructive and destructive findings, you could uncover an error in your revealed article. Whether or not the error is so simple as a typo or misplaced decimal level, or as drastic as incorrectly analyzed or introduced knowledge, prioritize its correction to keep away from the irreproducibility of the findings. 

Even when this implies retracting your article, it’s higher to have the error corrected than to have it proceed to misguide different researchers into performing irreproducible analysis. Nobody enjoys being mistaken or explaining how they’ve failed. However to enhance your analysis reproducibility, and the reproducibility of researchers who construct upon your findings, you should definitely embrace destructive knowledge and proper publication errors.

8. Be a Aware Researcher

Even when all the things else units the stage for a profitable and reproducible experiment, variations in researcher conduct may influence outcomes or the power to carry out experiments reproducibly.

For instance, are you too hopped up on caffeine to pipette precisely? Conversely, are you too drained to focus and keep away from errors? Or perhaps you’re distracted by issues happening outdoors of labor.

Discovering the suitable psychological area to carry out your experimental steps precisely and exactly is essential, and never one thing to be ignored or taken without any consideration. Even an ideal toolbox is most useful within the arms of a aware employee. If you’d like your analysis to be reproducible, it’s necessary to not solely good your toolbox but additionally be a aware researcher whereas utilizing that toolbox.

Analysis Replication in Abstract

There you have got it! Eight actionable ideas that may instantly enhance the reproducibility of your analysis, and that may make you a greater scientist. That can assist you bear in mind these key factors, we’ve summarized them in a easy desk (see Desk 1).

Desk 1:

Whether or not you’re on the experimental design stage, in the course of performing your experiments, or have already collected your outcomes and are looking for to publish your findings, the following pointers may help be certain that you proceed to make use of, construct, and enhance upon the truth-seeking toolbox of science.

Acquired any recommendation on making science extra reproducible? Share them within the feedback part!

References

1. Suresh Okay (2011) An outline of randomization methods: An unbiased evaluation of consequence in scientific analysis. J Hum Reprod Sci 4(1):8–11
2. Bespalov A, Wicke Okay, and Castagné V (2019) Blinding and Randomization. In: Bespalov A, Michel M, and Steckler T (eds) Good Analysis Follow in Non-Medical Pharmacology and Biomedicine. Handbook of Experimental Pharmacology, vol 257. Springer, Cham