Team:Tsinghua-A/Poster





Poster: Tsinghua-A



Poster

Here's our poster!

Our Poster
  The poster is made by our teammates, and it shares the same design style with our wiki and videos!

References

[1] Erlich Y, Zielinski D. DNA Fountain enables a robust and efficient storage architecture[J]. Science, 2017, 355(6328): 950-954.
[2] https://www.health.harvard.edu/cancer/radiation-risk-from-medical-imaging
[3] http://ai.stanford.edu/~gokturkb/Compression/FinalReport.htm
[4] Coupé C, Oh Y M, Dediu D, et al. Different languages, similar encoding efficiency: Comparable information rates across the human communicative niche[J]. Science Advances, 2019, 5(9): eaaw2594.
[5] Erlich Y, Zielinski D. DNA Fountain enables a robust and efficient storage architecture[J]. Science, 2017, 355(6328): 950-954.
[6]Trabelsi A, Chaabane M, Ben-Hur A. Comprehensive evaluation of deep learning architectures for prediction of DNA/RNA sequence binding specificities[J]. Bioinformatics, 2019, 35(14): i269-i277.
[7] Identifying enhancer–promoter interactions with neural network based on pre-trained DNA vectors and attention mechanism[J]. Bioinformatics, 2019.

Acknowledgement

Special Thanks to Zhe Zhao MD. who provides us help in HP section!
Introduction
  We are trying to put medical data into the brand new way of storing information ———— DNA. By storing information with DNA, we hope to throw light on the current system of medical data managing, which cannot provide both privacy and convenience for the patient.
  Our ultimate goal is to build a large system for whole population on whole dimension within whole life. But it is obvious that this is not gonna happen in near future. So we start from the ABCs. We designed a basic system and strage structure for dealing with these information and encoding algorithms for binary data.
Background——About DNA Storage
  Why are we using DNA storage? Because it has so many advantages! Compared to digital storage, it has much larger infomation density. Also, DNA storage provides much better information stability. DNA can store inofmation for centuries. They also use less energy compared to digita storage.
  In short, DNA storage is a storage method for the future!
Background——Medical Data
  Medical data involves a lot of things. It has much to do with a person's privacy. Also, medical data is a kind of "cold" data, which means it fits the feature of DNA storage!
Design——System Structure
  We designed a system structure for dealing the data, as well as storage system. Go to our Wiki pages for more information!
Design——Feature Extracting VAE
  We combine Artificial Neural Networks with DNA storage. For the medical image data set, the feature of the images in the data set are extracted by VAE, a powerful tool for projecting high dimension input into lower dimension feature space in machine learning area, and the primer are synthesized according to the obtained features.
  Through such a coding method, we can achieve the effect of "the more similar the original image, the more similar the sequence generated". The primers designed in this way are also the basis of our data retrieval and the realization of many functions in our project. Since finding the best match DNA is a chemical-like reaction and trying to find the best match in the digital world need be done one by one, this procedure thus show great promise once the database gets extremely large.
Engineering——Image Encoding
  Image is an important part in medical data and they take up lots of storing space. But actually doctors some time don't just need such clear images to diagnose. So we design a method which can saves up to 90% of storing space for most of the situations. This can save lots of space in the DNA chains.
Engineering——Waveform Encoding
  In order to facilitate a large number of ECG data storage, we designed a kind of compression method using the differential coding and variant-length coding algorithm of 2-4-8-16 bits. We coded demo in MATLAB and tested on MIT-BIH Arrhythmia Database. The results showed that our method can compress the raw data to about 20% of the original size, smaller than ZIP and RAR!
Engineering——Text Encoding
  In modern times, medical data continues to grow explosively. We urgently need to establish a huge medical database in order to realize the scientific management of hospitals and improve the level of medical services. Written medical record data is a very important part of medical data. It conforms to the characteristics of cold data and has long-term application value. So we hope to use DNA to store written medical records, so as to establish a patient's medical record database. If this idea is realized, at the individual level, it will be able to achieve accurate personalized medical care, make medical data never outdated, and in the medical industry, it will be able to obtain more transparent information in the data network, which has great value from medical services to drug research.
Human Practices——Medical Data's Future
  With the development of AI technology, the research value of medical data is growing rapidly. Therefore, how to store these medical data has become a new problem. Fortunately, DNA storage is just right for doing this. DNA storage's features can solve the existing problems in medical data storage right away.
  We connect medical data, DNA storage and blockchain. In Human Practices section, we discuss with teachers the possibility of DNA storage to realize medical data blockchain and a good conclusion has been obtained.