Blog
Sample Preparation, Mathematical Models, and Baffling Biology
Tiny flatworms may hold the key to stem cell technology, and precision sample preparation coupled to mathematical modeling may be the way to unlock it.
April 14, 2022
Planarians are diminutive freshwater flatworms that are perhaps best known among biologists (and high school students) for their unique regenerative abilities. So far as we know, there is no other organism capable of such a complete repair and regrowth of damaged or missing parts. Planarians have been documented regenerating an entire, complete, and functional body from just 1/200th of their body, roughly the equivalent of a human regrowing their entire body from just three or four fingers. A decapitated planarian will spawn two new organisms, with the head growing a new tail and amazingly the tail growing a fully functional head, nervous system, eyes, and all. This phenomenal ability to regenerate is driven by the planarian’s stem cells, which remain functional in the adult worm, and in recent years this fact has garnered the interest of scientists as a model of stem cell biology in action. On top of all this, planarians share much in common with higher organisms including vertebrates and even humans, using the same or similar neurotransmitters and sharing many complementary genes. All of which makes planarians a tantalizing model and a steppingstone towards stem cell therapies.
Navigating the path to stem cell therapies
However, like every good story, there is a catch that means it’s not all just plain sailing. While planarian biology of has been studied extensively, a good deal less is known about its genetics and protein expression, details that are extremely important when figuring out which genes are related to planarian regeneration, and how those genes function.
Planarians are what researchers refer to as a ‘Non-Model Organism’ which means that they haven’t been studied by the research community to the same extent as common research organisms such as fruit flies or zebrafish. Studies going back for decades mean that we know these model organisms inside out, we know the layout of their genome in detail. Not so for a non-model organism such as planarian.
Non-model organism: Too much data, not enough information.
RNA analysis of non-model organisms is often difficult because their genome isn’t well annotated. Much less is known about where genes are located in relation to each other, or which genes may be turned on or off under different conditions.
Examining differentially expressed transcripts (DETs) is an important method for understanding genetic changes during disease onset or progression. Changes in protein expression can be linked to distinct biological processes, and also to the genes involved in those processes. The identification of DETs in a non-model organism, however, is exceedingly difficult. It can be unclear whether short genetic sequences overlap, how much redundancy is present, and whether or how often sequences are transcribed. There is simply too much data and too little annotation.
Mathematical modeling to reduce data complexity
In a recent publication[1] a team at Gakuin University in Tokyo proposed a strategy to rapidly parse the planarian genome and identify key genes involved in regeneration and the control of stem cells. Coupling together high-quality preparation of biological samples from two planarian populations with complex mathematical modeling. In the last few years, mathematical techniques designed to simplify computer modeling and data mining have increasingly been used to unravel complex biological systems.[2] In this study, the scientists use a technique known as ‘Tensor Decomposition-Based Unsupervised Feature Extraction’.
While the technique is a mouthful, the premise behind it is reasonably straightforward. Tensor decomposition (TD) is a way to take a process with multiple variables, such as those involved in gene expression, and separate out each variable so that the data becomes more manageable for processing. Feature extraction (FE) is used to reduce data redundancy. It uses pattern recognition to identify key features and patterns in the data. For example, one could examine only those gene transcripts which change over time, or under different experimental conditions.
Sample preparation ensures reliable results
To test their new mathematics-based strategy, the researchers first needed a high-throughput way to obtain high-quality RNA sequences from planarian. Clean data is essential to the success of the mathematical modeling since poor resolution or the presence of contaminants will confound an already convoluted data set, creating artifacts and noise that can throw off the algorithm.
The primary goal of this study was to see whether the new deconvolution technique could be used to distinguish gene transcripts involved in normal versus defective planarian regeneration. To accomplish this, the scientists used RNAi technology to systematically knock down individual gene transcripts to help determine whether that transcript contributed to a loss of regenerative function.
The authors used AcroPrep™ Advance 96-well long tip filter plates to extract RNA from individual planarian in the study. AcroPrep plates are optimized for robotic and automated handling, and specifically constructed to minimize the risk of sample cross-contamination, both key attributes required by the protocol in order to generate high numbers of samples in a fully automated fashion, and to eliminate the possibility of noise from contamination.
Evaluation of the biological data proved to be challenging, but the results from the study were encouraging. The researchers successfully identified a subset of 155 transcripts whose expression changed over time and was differentially expressed between normal planarians and those with defective regeneration. Effectively demonstrating the utility of TD-based FE analytical techniques for the elucidation of genes of interest in hitherto poorly characterized non-model organisms and helping to pinpoint the genes and pathways of interest to researchers.
Pall’s AcroPrep Advance filter plates are available for a wide range of applications and sample volumes. You can learn more about the AcroPrep Advance Filter Plates used in this study on the Pall website.
References
1. Kashima M., et al. RNA-Seq data analysis for Planarian with tensor decomposition-based unsupervised feature extraction. bioRxiv (2021) [preprint] June 2021 https://doi.org/10.1101/2021.06.15.448531
2. Ng, K-L and Taguchi Y. Identification of miRNA signatures for kidney renal clear cell carcinoma using the tensor-decomposition method. Nature Scientific Reports (2020) 10 (151490) September 2020 https:// doi.org/10.1038/s41598-020-71997-6