Team:Peking/Poster

Poster: Peking

Music Generation, Updating and Visualization Based on DNA Storage and Gene Editing

Nan Huang¹, Kai Gao¹, Zhuoran Li², Yiliang Chen², Anqi Yu¹, Meijie Pan³, Letong Wang³, Yuxuan Wu², Ruoxuan Zhuang³, Shangkun Wang¹, Long Zhao², Xuelong Chen³, Ze Cao³

¹School of Life Science, Peking University, ²Yuanpei College, Peking University, ³Peking University Health Science Center

Abstract

Based on DNA storage and gene editing, we constructed a prokaryote system that realizes our goal of storage, evolution and visualization of music, combining science with art. First, we designed a coding method that stores music information in DNA. After that, we introduced chemically synthesized DNA sequence encoding music into E. coli, mutated it with base editors and EvolvR, and extracted music from altered sequence. To complete our music evolution system, we designed a series of scoring functions that quantify the music's harmony, which acts as the selection pressure. Besides, we adopted similar strategy to convert DNA sequence into images. For further promotion of our idea, we held an exhibition themed the beauty of life and won applause from the public.

Introduction

Due to its stability and high storage density, DNA is a promising storage matrix. It is estimated that DNA can be stably stored for more than 100 years, and 215PB of data can be stored in one gram of DNA[1]. Since 2010s, information including program scripts, literacy, images, music and videos, has been successfully written into DNA and extracted[2].

Genetic algorithm is a widely used method in computer science to evolve information, e.g. obtaining more beautiful music from original music through evolution. Life shows random patterns that differs from any computer-simulated randomness, providing more possibilities for information evolution. However, currently there lacks methods compatible with genetic algorithm-like information evolution for DNA storage of music and images.

Gene editing is a burgeoning field which offers powerful tools to induce mutations such as transition, transversion, indel or recombination in pre-existing DNA sequence in vivo[3]. Here, we show that with well-designed coding algorithms, gene editing can be applied to plasmid or genome encoding art pieces to alter and update the encoded information. We provide key modules for generation, update and visualization of music encoded in plasmid in E. coli.

References

[1] Dong, Y., Sun, F., Ping, Z., Ouyang, Q., & Qian, L. (2020). DNA storage: research landscape and future prospects. National Science Review. 7(6):1092–1107.

[2] Ceze, L., Nivala, J., & Strauss, K. (2019). Molecular digital data storage using DNA. Nature Reviews Genetics. 20:456–466.

[3] Anzalone, A.V., Koblan, L.W. & Liu, D.R. (2020). Genome editing with CRISPR–Cas nucleases, base editors, transposases and prime editors. Nature Biotechnology. 38:824-844.

Goal & Design

At the very beginning of our project, we have set our goals to store music in DNA and perform mutation both in silico and/or in vivo, and of course visual art pieces as well. The first step in our designed workflow is to figure out a universal rule to encode music into DNA sequence and decode music from DNA sequence. Then, we need to perform mutation in DNA, and we would evaluate how good it is and convert the better ones back to DNA for next round of mutation. Besides, we would like to create visual art from DNA sequences and is kept compatible with the rules of converting music into DNA to reach harmony between audio and vision.

Coding & Analysis

We designed some algorithms to encode music into DNA sequence and decode music from DNA sequence. There are 4 versions of decoding algorithms. Their sketch, pros and cons are listed below.

Generation	How to generate music	Pros	Cons
Pitch-based system	Every 3 bases for a sampling point (a semiquaver)	It’s simple.	The music is simple, and most music become noise after random mutation.
Instruction-based system	Operate on single note of original music composition for every 6 bases, to alter pitch and duration	More powerful than Encode 1.0; Flexible for encoding and directed evolution	Occasionally produces unpleasing tunes
Enhanced Instruction-based system	Operate on some notes in a piece for some number of bases, to alter pitch and duration	Exaggerates effect of mutations	Occasionally produces unpleasing tunes; too complicated to encode music into DNA
Music theory-based system	DNA sequence is mapped onto a hierarchal rule set, based on music theory	Provides pleasant tunes	Too complicated to encode music into DNA

To enable automatic evolution of DNA segments that encode music pieces, we designed a scoring algorithm (Fig.A) which exerts selection pressure on descendants of original DNA sequence and picks the sequences encoding high-score music (Fig.B).

Experiment & Result

Visualization

According to an article published by Church in 2017[1], there are two basic logics for using bases to encode images. One is "one-one mapping". In simple terms, it is to use one base to correspond to one color. The advantage of this method is that it is simple and not easy to make mistakes. One more or one less oligo has less impact on the encoding. The shortcomings are obvious, the color and form are too monotonous, and the GC content cannot be controlled. The other method is relatively complicated. To formulate a "color codon table" similar to a "codon table", specifically, you must first formulate a coding rule, using a few bases to represent a color. The advantage of this method is that it is helpful for error correction, and it can design codon similar to the stop codon. It provides a strong guarantee for the stability, logic and diversity of coding rules.

Here we present three visualizations, other types and specific methods available on our wiki. The three categories are elements combination, the art of “splicing” and ink bamboo. In elements combination, regular geometric elements or non-geometric elements are placed on the canvas according to certain rules. In the art of “splicing”, we used geometric figures to represent different bases and stitched them together closely on the canvas. In ink bamboo, we combined traditional Chinese art with DNA. First, we simulate the handwriting of the brush, then specify the rules that will reflect the genetic code in the ink painting.

Fig. Visualization art a). The art of “splicing”. b). Elements combination. c),d). Elements combinations with specific backgrounds. e). Ink bamboo

References

[1] Shipman SL, Nivala J, Macklis JD, Church GM. (2017) CRISPR-Cas encoding of a digital movie into the genomes of a population of living bacteria. Nature. 547(7663):345-349.

Acknowledgement

GENERAL SUPPORT

Prof. Wei Ping, Prof. Ouyang Qi, Prof. Qian Long

EXPERIMENT EQUIPMENT SUPPORT

Prof. Wei Ping

School Department

	PEKING UNIVERSITY HEALTH SCHOOL CENTER
	YUANPEI COLLEGE PEKING UNIVERSITY
	Center for Quantitative Biology, Peking University
	School of Life Sciences, Peking University
	Office of Educational Administration, Peking University

Institutes, Organizations and Facilities

	Central Academy of Fine Arts
	Bluepha
	Shenzhen Insititutes of Advanced Technology Chinese Academy of Sciences