Take my Breath Away: Diagnosing Asthma’s Severity When Every Minute Counts

Take my Breath Away: Diagnosing Asthma’s Severity When Every Minute Counts

When I first came to the Colorado, the mountains captivated me. They looked so imposing, yet enchantingly beautiful at the same time. A few months later, some of my mountaineering friends convinced me to climb one of these enchantresses (known locally as “14’ers”). They picked an “easy” one because I had spent all my life at a few hundred feet above sea level near the Mississippi River.

We started out on a beautiful crisp fall morning. The sun had yet to rise and illuminate the aspen trees that were already turning a golden yellow. As we climbed higher, the aspen grew smaller and it got colder. As the air grew colder and thinner, I found myself having a difficult time breathing. I kept soldiering on, but the fight for breath was getting harder. According to one friend’s watch, I had made it about 13,700 feet before the struggle for breath grew too much. I had to retreat back to a more oxygen-rich environment. The mountain won this round.

There are individuals who experience the fight for breath every day. The cause of this fight is a condition known as asthma, which can vary greatly in its severity. Diagnosing the severity and determining the correct course of treatment is not always straightforward, quick or cheap (Israel & Reddel, 2017). If new diagnostic tests became available, could they speed up the process of determining asthma severity and thus identifying the best treatment?

An international team of researchers united to answer that very question (Rossios et al., 2017). They queried sputum (another name for phlegm) samples from patients with different degrees of asthma to look for changes in the patients’ transcriptomic (looking at all RNA levels) and proteomic (looking at all protein levels) profiles. The researchers successfully found changes in those profiles that provide new insights about the underpinnings of asthma severity and may even help expedite the diagnosis (Rossios et al., 2017).

These researchers took not just one small step, but one giant leap towards summiting Mt. Improved Diagnostics. Instead of focusing on just one biomarker and looking for its presence in samples provided by patients with different degrees of asthma severity, the researchers utilized technologies that could cast a broad net (Rossios et al., 2017). Using the SOMAscan assay, they could scan various molecular pathways simultaneously, see the differences and achieve a better understanding (Rossios et al., 2017).

The asthma researchers certainly share a great vantage point with others who use proteomics. Proteins, which are the end product of our genes, are responsible for how our bodies respond to the environment, disease, etc. Aside from responding to cues, rogue proteins can also be the cause of disease. By looking at how proteins interact with one another and the downstream effects of those interactions, the scientific community can better discern the onset of disease (Fessenden, 2017). By thinking deeply about the data, it will feasible to scale the enchantress, Mt. Improved Diagnostics, with greater ease and surer breath.



Fessenden, M. (2017). Protein maps chart the causes of disease. Nature, 549(7671), 293-295. doi:10.1038/549293a

Israel, E. & Reddel, H. K. (2017). Severe and difficult-to-treat asthma in adults. N Engl J Med, 377(10), 965-976. doi:10.1056/NEJMra1608969

Rossios, C., Pavlidis, S., Hoda, U., Kuo, C. H., Wiegman, C., Russell, K., . . . Unbiased Biomarkers for the Prediction of Respiratory Diseases Outcomes Consortia Project Team. (2017). Sputum transcriptomics reveal upregulation of IL-1 receptor family members in patients with severe asthma. J Allergy Clin Immunol. doi:10.1016/j.jaci.2017.02.045


A Beautiful Brain Besieged by Glioblastoma

The frog lies there, all splayed out and pinned. It will no longer ribbit or hop. Gone are its chances of being transformed into a prince with a simple kiss. Yet, this formaldehyde-soaked creature still fascinates. For it reveals to young eyes caught up in an anatomy lesson, just how intricate and miraculous biological bodies truly are.

Take the exposed brain peeking at the dissector through a hole in the frog’s cranium, for instance. Translucent lobes glisten in the fluorescent light. These lobes and other brain parts are the locus for everything that makes us (and even the frog) who we are. It’s the most terrifying of circumstances when a person is robbed of their very identity by brain cancer.

Glioblastoma, which has been in the news recently (both in politics and sports), is a particularly nasty form of brain cancer. It is highly aggressive, rapidly pillaging a person’s identity. It comes without any significant warning, perhaps via generic symptoms like a headache or nausea (Geez. This sounds like how a lot of things start.) (Young, Jamshidi, Davis, & Sherman, 2015). Victims can also experience a change in personality or memory loss pending on where in the brain the cancer is located (Young et al., 2015). Even with treatment, the average victim can fend off the pillager for only a little while (Davis, 2016; Young et al., 2015).

New and better treatments are obviously needed, which in turn require a better understanding of this plunderer. In an elegant assay, researchers demonstrated that the environment (not the genes) dictates the cancer’s pilfering path (Miller et al., 2017). Pending the testing (environmental) conditions, different proteins that regulate the making of RNA (a message for making protein) are activated. This in turn can affect how responsive the tumors are to certain chemotherapies (Miller et al., 2017). Given how these proteins (i.e., transcription factors) wielded such a huge impact on how the tumor responded to its environment, this work suggests that transcription factors are logical targets for new therapeutics.

It is still too soon to tell if this work will result in new therapeutics that safeguard our repository of uniqueness. We can be cautiously optimistic, but other research may burst that beautiful bubble (Mak, Evaniew, & Ghert, 2014). More and more evidence continues to show that what may save the life of lab animals will not work for us experimenters (Ugh).

So, what are we to do? Sit around, look pretty and twiddle our thumbs? No! We need to rise and emit some battle cries. We must guard that optimistic bubble. New procedures are needed to improve the translation of successes in the lab to the clinic. We also need to continue developing new defenses against glioblastoma. If we can develop better sentries (diagnostics), it is possible to spot the marauding tumors sooner. Even Dr. Philip E. Steig (founder and Chairman of Weill Cornell Brain and Spine Center) shares in this optimism that earlier detection, improved knowledge and new treatments could improve the odds of the cancer going into remission (Steig, 2016). The sooner we can tell if that headache is nothing versus cancer, the sooner we can battle and hold onto what constitutes our (and the frog’s) beautiful individuality.


Davis, M. E. (2016). Glioblastoma: Overview of Disease and Treatment. Clin J Oncol Nurs, 20(5), S2-8. doi:10.1188/16.CJON.S1.2-8

Mak, I. W., Evaniew, N., & Ghert, M. (2014). Lost in translation: animal models and clinical trials in cancer treatment. Am J Transl Res, 6(2), 114-118.

Miller, T. E., Liau, B. B., Wallace, L. C., Morton, A. R., Xie, Q., Dixit, D., . . . Rich, J. N. (2017). Transcription elongation factors represent in vivo cancer dependencies in glioblastoma. Nature, 547(7663), 355-359. doi:10.1038/nature23000

Young, R. M., Jamshidi, A., Davis, G., & Sherman, J. H. (2015). Current trends in the surgical management and treatment of adult glioblastoma. Ann Transl Med, 3(9), 121. doi:10.3978/j.issn.2305-5839.2015.05.10

Steig, P.E. (2016, August 8). Early Detection Can Be Key to Surviving a Brain Tumor. Retrieved from http://weillcornellbrainandspine.org/early-detection-can-be-key-surviving-brain-tumor.


A Mermaid’s Tale About Knowledge Sourced from Genomics

A Mermaid’s Tale About Knowledge Sourced from Genomics

In Disney’s The Little Mermaid (1989), a vivacious, inquisitive mermaid becomes infatuated with humans. She eagerly learns everything she can about these mysterious creatures. In her scholarly pursuits, she talks extensively with a seabird who satiates her curiosity with extremely inaccurate knowledge, such as using a fork to comb hair. While we might chuckle at the misuse of the fork, we too often apply knowledge that only in hindsight causes chuckling or a quiet reflection of “OMG! What the *@#% were we thinking?”

For instance, consider genomics. Upon the near completion of the Human Genome Project, shell trumpets and fish choruses heralded the genetic revolution that was going transform medicine. In 2001, Francis Collins (current Director of the National Institutes of Health) and Victor McKusick co-wrote, “Genomic medicine holds the ultimate promise of revolutionizing the diagnosis and treatment of many illnesses (Collins & McKusick, 2001).” Since the completion of the project, it seems the floodgates have opened, releasing waves of genetic tests.

Nearly two decades later, the wave of enthusiasm is crashing. In a recent survey, most oncologists acknowledged that genomic testing does not meet expectations and has been overhyped (Genomeweb, 2017). They and other medical professionals also admitted that they did not have the expertise to adequately decipher the large amount of genetic data or to communicate them effectively to patients (Genomeweb, 2017; (Mikat-Stevens, Larson, & Tarini, 2015). Doctors often relied on the test manufacturer’s interpretations (Graber, 2015).

Consequently, knowledge acquired from genetics is not applied correctly. “Poor unfortunate souls” are getting over diagnosed and/or receiving unnecessary treatment based on the results. Recently, a young teenager received an implantable defibrillator based on genetic tests that said he was at risk for a fatal heart condition (Ackerman et al., 2016). A second opinion revealed that he did not have the fatal syndrome and that the implantable defibrillator was not necessary (Ackerman et al., 2016). Another study highlighted that half of women who undergo double mastectomies in the hopes of avoiding breast cancer had mutations, but not the kind known to increase cancer risk (Conger, 2017). Needless to say, a significant percentage of the doctors said that they would treat these women with mutations of unknown significance the same as those with the mutations known to increase risk (Conger, 2017).

It would have been ideal to have a clinical geneticist help guide both doctor and the patient through the murky data. However, there are too few of them for the huge demand placed upon them (Graber, 2015). Even some admit that they are having a hard time keeping up with the literature and new findings (Graber, 2015), which can totally change how one evaluates risk of disease based on genetics (Boyle, Li, & Pritchard, 2017; Chen et al., 2016).

Our understanding of genetics is really a drop in the ocean. As mentioned earlier, new findings are highlighting how little the scientific and medical communities know about a “simple” code. Healthy people are walking about with what might be considered “bad” genetics. Through lifestyle choices or some other undiscovered biological reason, these people are just fine (Chen et al., 2016; Khera et al., 2016). Callouts are being made to change how genetic contributions to diseases are determined (Boyle et al., 2017). Let’s face it, genomes are more complex than anyone could have foreseen.

Have we come to a point where we start quietly reflecting about the use of genomics in medicine? Yes. Already, the National Cancer Institute and the US Food and Drug Administration are considering proteogenomics as the new frontier for clinical diagnostics (Bonislawski, 2017). This approach, which couples genetic information with protein information to improve understanding of biological processes, could yield a more favorable outcome warranting celebratory singing by a Caribbean crab with a thick accent. As to not repeat the past, however, extreme care should be made to ensure that results and their limitations are clearly communicated and recipients know how to properly use the knowledge.


Ackerman, J. P., Bartos, D. C., Kapplinger, J. D., Tester, D. J., Delisle, B. P., & Ackerman, M. J. (2016). The Promise and Peril of Precision Medicine: Phenotyping Still Matters Most. Mayo Clin Proc. doi:10.1016/j.mayocp.2016.08.008

Bonislawski, A. (2017, July 21). FDA and NCI Memorandum Indicates Growing Interest in Proteogenomics as

Clinical Approach. Retrieved from https://www.genomeweb.com/proteomics-protein-research/fda-nci-memorandum-indicates-growing-interest-proteogenomics-clinical.

Boyle, E. A., Li, Y. I., & Pritchard, J. K. (2017). An Expanded View of Complex Traits: From Polygenic to Omnigenic. Cell, 169(7), 1177-1186. doi:10.1016/j.cell.2017.05.038

Chen, R., Shi, L., Hakenberg, J., Naughton, B., Sklar, P., Zhang, J., . . . Friend, S. H. (2016). Analysis of 589,306 genomes identifies individuals resilient to severe Mendelian childhood diseases. Nat Biotechnol, 34(5), 531-538. doi:10.1038/nbt.3514

Collins, F. S., & McKusick, V. A. (2001). Implications of the Human Genome Project for medical science. JAMA, 285(5), 540-544.

Conger, K. (2017, April 12). Physicians’ misunderstanding of genetic test results may hamper mastectomy decisions for breast cancer patients. Retrieved from https://med.stanford.edu/news/all-news/2017/04/misunderstanding-genetic-test-results-may-increase-double-mastectomies.html

Genomeweb (2017, May 2). In Survey, Oncologists See Genomic Testing as Important Advance, But Value

‘Below Expectations.’ Retrieved from https://www.genomeweb.com/molecular-diagnostics/survey-oncologists-see-genomic-testing-important-advance-value-below.

Graber, C. (2015, February 5, 2015). The Problem with Precision Medicine. The New Yorker.

Khera, A. V., Emdin, C. A., Drake, I., Natarajan, P., Bick, A. G., Cook, N. R., . . . Kathiresan, S. (2016). Genetic Risk, Adherence to a Healthy Lifestyle, and Coronary Disease. N Engl J Med, 375(24), 2349-2358. doi:10.1056/NEJMoa1605086

Mikat-Stevens, N. A., Larson, I. A., & Tarini, B. A. (2015). Primary-care providers’ perceived barriers to integration of genetics services: a systematic review of the literature. Genet Med, 17(3), 169-176. doi:10.1038/gim.2014.101

A Mad Hatter’s Question About Correlating Transcriptomics to Proteomics

A Mad Hatter’s Question About Correlating Transcriptomics to Proteomics

Why is a raven like a writing desk? Lewis Carroll penned this head-scratcher over a century ago in his book, Alice’s Adventures in Wonderland. Since then, people keep trying to draw parallels between these two unrelated items to answer the Mad Hatter’s baffling riddle. In Omicsland, a similar riddle could be uttered at a social gathering, “Why is transcriptomics (looking at the RNA levels in an individual) like proteomics (looking at the protein levels in an individual)?”

If logic based on biology’s central dogma (DNA begets RNA which begets protein) is applied, the initial response might be because they are equivalent with respect to tracking protein levels. If the level of mRNA (messenger RNA that codes for the protein) rises, then the amount of the corresponding protein would rise too. This logic is not entirely sound. It was shown that mRNA levels do not always correlate with protein levels (Vogel & Marcotte, 2012). Recently, additional research has poked holes into research saying that the two omics correlate well (Fortelny, Overall, Pavlidis, & Freue, 2017). I’ll briefly elaborate about this lack of correlation, but a more thorough explanation (Liu, Beyer, & Aebersold, 2016) bears a tag saying, “Read me.”

One mRNA does not beget just one protein. Many proteins can be created from the same mRNA. The efficiency of this process can be affected by several factors. One of the major influences can be found within some, but not all mRNAs themselves. The mRNA can possess chemical modifications that can affect the process of making protein (Zhao, Roundtree, & He, 2017), possess internal elements that serve as video game-like cheat codes to fast track the process (Walters & Thompson, 2016) or contain binding sites for proteins that help regulate when the mRNA should be used (Nelson, Leidal, & Smibert, 2004). Another major influence can be found in the regulation of proteins (besides the ribosome) that are involved in converting the mRNA code into a protein (Nho & Peterson, 2011).

Aside from biological reasons, technology issues can sometimes explain why mRNA levels do not correlate with protein levels. Variations in how a technique is executed and how data are analyzed abound, and can affect the results. Also, technical approaches have their limits and may not be the best ones to use for certain tasks (e.g., using flamingos as croquet mallets). Best practices and new approaches are being proposed to help address the limits and reduce the variation that can arise (Conesa et al., 2016; Hu, Noble, & Wolf-Yadlin, 2016).

As noted earlier, correlation between proteomics and transcriptomics is low. However, a small percentage of protein levels do correlate with mRNA levels. This correlation, however, may only happen in certain instances or biochemical pathways (Liu et al., 2016; Zhang et al., 2016).

Let’s revisit the original question, why is transcriptomics like proteomics? The answer could simply be that the two are alike because they are both complicated. Correlating the two is feasible, but not without peril. With a low correlation, is it worth jumping down the rabbit hole to draw parallels between the two?


Conesa, A., Madrigal, P., Tarazona, S., Gomez-Cabrero, D., Cervera, A., McPherson, A., . . . Mortazavi, A. (2016). A survey of best practices for RNA-seq data analysis. Genome Biol, 17, 13. doi:10.1186/s13059-016-0881-8

Fortelny, N., Overall, C. M., Pavlidis, P., & Freue, G. V. C. (2017). Can we predict protein from mRNA levels? Nature, 547(7664), E19-E20. doi:10.1038/nature22293

Hu, A., Noble, W. S., & Wolf-Yadlin, A. (2016). Technical advances in proteomics: new developments in data-independent acquisition. F1000Res, 5. doi:10.12688/f1000research.7042.1

Liu, Y., Beyer, A., & Aebersold, R. (2016). On the Dependency of Cellular Protein Levels on mRNA Abundance. Cell, 165(3), 535-550. doi:10.1016/j.cell.2016.03.014

Nelson, M. R., Leidal, A. M., & Smibert, C. A. (2004). Drosophila Cup is an eIF4E-binding protein that functions in Smaug-mediated translational repression. EMBO J, 23(1), 150-159. doi:10.1038/sj.emboj.7600026

Nho, R. S., & Peterson, M. (2011). Eukaryotic translation initiation factor 4E binding protein 1 (4EBP-1) function is suppressed by Src and protein phosphatase 2A (PP2A) on extracellular matrix. J Biol Chem, 286(37), 31953-31965. doi:10.1074/jbc.M111.222299

Vogel, C., & Marcotte, E. M. (2012). Insights into the regulation of protein abundance from proteomic and transcriptomic analyses. Nat Rev Genet, 13(4), 227-232. doi:10.1038/nrg3185

Walters, B., & Thompson, S. R. (2016). Cap-Independent Translational Control of Carcinogenesis. Front Oncol, 6, 128. doi:10.3389/fonc.2016.00128

Zhang, H., Liu, T., Zhang, Z., Payne, S. H., Zhang, B., McDermott, J. E., . . . Investigators, C. (2016). Integrated Proteogenomic Characterization of Human High-Grade Serous Ovarian Cancer. Cell, 166(3), 755-765. doi:10.1016/j.cell.2016.05.069

Zhao, B. S., Roundtree, I. A., & He, C. (2017). Post-transcriptional gene regulation by mRNA modifications. Nat Rev Mol Cell Biol, 18(1), 31-42. doi:10.1038/nrm.2016.132

Understanding the Ramifications of Mutations: Seeing Past the Trees to the Forest

Understanding the Ramifications of Mutations: Seeing Past the Trees to the Forest

Remember Jurassic Park?  I can. I still get shivers thinking about the dinosaurs crashing through the trees and devouring anything that moves. I also recall the adorable animation the park visitors watched to better understand how the dinosaurs were brought back from extinction via the DNA in their blood cells trapped in those ancient mosquitoes encased in amber. This introduction to genetic engineering via Hollywood initiated my scientific endeavor for learning more about nucleic acids. I was hooked.

Since then, I have joined the countless masses who have become enthralled with the double-helix. We believed that the message that dictated our uniqueness resided in that simple, repetitive molecule. Thanks to technological advances, reading these messages has become both easy and affordable. As a result, the scientific literature is inundated with genome wide association studies (GWAS), which search for genetic changes that are related to many different diseases and conditions (Manolio, 2017).

Boyle et al. raised the question as to whether or not they should limit their focus to just a small subset of genes identified from GWAS to better understand disease or inheritable traits (Boyle, Li, & Pritchard, 2017). In their recent Cell publication, the group of scientists concluded that while a “core” set of genes might be a cause for something, the overall phenotype (physical or observable characteristics) resulted from many small contributions from many more genes. The authors stated that while the initial impulse may be to do even more GWAS, a different approach may be needed. They recommended doing an omnigenics analysis of GWAS data.

What is “omnigenics” you may ask? For complex traits or diseases, any gene variant in the genome could be contributing to the manifestation of the disease or trait. Hence, to truly understand the genetic cause of a disease or trait, all DNA mutations need to be considered. Even if the mutation seems to affect a system seemingly unrelated to the disease or trait of interest, it should be included in the analysis.

Recently, a study was published that showed that individuals were carrying disease-causing gene variations, yet the carriers showed no physical manifestation of the disease (Chen et al., 2016). The reason could be what the Boyle et al. had suggested: The manifestation is dictated by the genomic forest and not by a few gene trees.

This omnigenics approach sounds lovely, but it comes with its own issues. How do we truly separate legitimate changes from the background to understand the causes of disease? If we look at the data, is it possible that we are just looking at a pile of sawdust trying to pretend to be the desired forest? With so much information embedded in DNA, it can be a daunting challenge to establish causality no matter how good the software that sifts through the info.

Instead of considering our DNA, maybe it would be worthwhile to change focus and look at the proteomic (all proteins that constitute our bodies) forest. Being the products of our genes, proteins offer a timelier view of what is happening in our bodies. Understanding how changes in the protein, such as concentration, correlate with disease or inheritable traits may shed more light onto how different biological systems interact with one another in the body. It could also help explain exactly how simple genetic change(s) in an unrelated pathway could affect a physical trait. This capability would certainly explain what happened in the Jurassic Park dinosaurs, which had frog DNA inserted into their genomes, to allow some of them to become male.


Boyle, E. A., Li, Y. I., & Pritchard, J. K. (2017). An Expanded View of Complex Traits: From Polygenic to Omnigenic. Cell, 169(7), 1177-1186. doi:10.1016/j.cell.2017.05.038

Chen, R., Shi, L., Hakenberg, J., Naughton, B., Sklar, P., Zhang, J., . . . Friend, S. H. (2016). Analysis of 589,306 genomes identifies individuals resilient to severe Mendelian childhood diseases. Nat Biotechnol, 34(5), 531-538. doi:10.1038/nbt.3514

Manolio, T. A. (2017). In Retrospect: A decade of shared genomic associations. Nature, 546(7658), 360-361. doi:10.1038/546360a

Cancer: The Ultimate Malware

A horrified astronaut utters over the radio, “Uh…Houston, we have a problem. Someone just hacked our computers. Now, we are viewing a message that states we must pay a ransom of 20,000 bitcoins or lose our top-20 karaoke playlist! Please advise.” I bet this conversation never occurred during lunar missions from the 60’s and 70’s.

Today, our beloved cell phones carry more computational capacity than the computers used to get men to the moon (NASA, 2017). As we become even more dependent on phones and other computers to help navigate our everyday lives, we become more vulnerable to malicious hackers or malware that can render them useless, or worse, steal valuable data.

Cancer, a biological equivalent to hackers and malware, can overtake our bodies and create such havoc that it disrupts our day-to-day lives or even ends them. As in the computer technology sector, large resources are being poured into figuring out how the hacks occur and how to remedy the situation. Recently in Nature, two articles were published detailing hacking methods used by some cancers that involve taking over how cells normally communicate with one another and control cell fate.

In a typical scenario, cells communicate with one another using proteins that decorate the outer surface of the cell or are excreted (Perrimon, Pitsouli, & Shilo, 2012). These proteins will bind to another protein (known as a receptor) found on the surface of another cell. This binding event triggers a cascade of internal events that can cause a cell to carry out a specific function, such as transforming into a different cell. Pending how the involved proteins have been modified (e.g., by sugars, phosphates or other chemical compounds added to the proteins), the resulting cascade can have very different outcomes. For instance, this type of cellular communication can tell an embryonic cell to become part of a hand, foot, heart, brain, etc.

When cancer hacks the system, normal cell fates are compromised. In lung adenocarcinoma for example, Tammela et al. found that the tumor cells can differentiate into two types (Tammela et al., 2017). One type is a typical tumor cell. The second cell type almost appears like a “normal” cell, but it is producing proteins that can fuel the cancer (think of adding gasoline to a raging fire). In another study, Lim et al. also saw how cancer cells can fuel their own fire (Lim et al., 2017). In small-cell lung cancer, neuroendocrine tumor cells, which produce hormones (messages to other cells) in response to signals received from the nervous system, switch to a different cell type upon activation of a pathway that can suppress tumor growth. These new cancer cells tend to be resistant to chemotherapy, and produce signals that encourage proliferation of the original neuroendocrine tumor cells. In these two studies, the authors suggest that these hacking strategies could be the source for new biomarkers or targets for new therapeutics.

As our understanding of this malicious hacker/malware improves, we can develop better diagnostics or patches (therapeutics) that can protect our most valuable asset, our health. How nice would it be to go to a doctor’s office, take a blood test and learn that we need the anticancer patch v2.0? This is already a reality for our phones and computers. Only time will tell if it becomes reality for the doctor office equivalent. If we can get a person to the moon with technology that fits (in most cases) in our back pocket, then maybe the time is getting closer?


Lim, J. S., Ibaseta, A., Fischer, M. M., Cancilla, B., O’Young, G., Cristea, S., . . . Sage, J. (2017). Intratumoural heterogeneity generated by Notch signalling promotes small-cell lung cancer. Nature, 545(7654), 360-364. doi:10.1038/nature22323

National Aeronautics and Space Administration (NASA). Do-It-Yourself Podcast: Rocket Evolution. Retrieved on June 17, 2017 at https://www.nasa.gov/audience/foreducators/diypodcast/rocket-evolution-index-diy.html

Perrimon, N., Pitsouli, C., & Shilo, B. Z. (2012). Signaling mechanisms controlling cell fate and embryonic patterning. Cold Spring Harb Perspect Biol, 4(8), a005975. doi:10.1101/cshperspect.a005975

Tammela, T., Sanchez-Rivera, F. J., Cetinbas, N. M., Wu, K., Joshi, N. S., Helenius, K., . . . Jacks, T. (2017). A Wnt-producing niche drives proliferative potential and progression in lung adenocarcinoma. Nature, 545(7654), 355-359. doi:10.1038/nature22334