"Borderlands" is a game series that combines first-person shooter and role-playing elements. In the latest installment, "Borderlands 3," there is a unique in-game game known as the "Sanctuary III" arcade on board the spaceship. In this arcade game, players are immersed in a retro-flavored mini-game where the challenge is to arrange blocks of different colors and patterns distributed across 7 to 20 columns. The primary task is to ensure that each column's patterned blocks align with the reference pattern on the far left by arranging specific gap blocks (as shown by the yellow arrow blocks), i.e., placing the same patterns in the same row.

In "Borderlands Science," achieving higher levels of alignment in the mini-game earns players higher scores and generous in-game currency rewards, up to several thousand tokens. These tokens can be used to buy various in-game skins and upgrade packs that enhance the player's capabilities in the main story.

This widely popular mini-game has been experienced by more than 4 million players to date. You might be surprised to learn that the data generated from this simple arcade-style mini-game, originating from a top-tier shooting game, has surpassed the most advanced computer algorithms in scientific research and has become a key technology in a study published in the well-known international journal "Nature Biotechnology."

In fact, every patterned block in the mini-game is not randomly arranged; each column represents the genomic sequences of real-life organisms on Earth—the gut microbiome's genetic sequences. Each pattern represents a specific DNA nucleotide (A, T, C, and G), with information sourced from the human microbiome research project under the Microsetta Initiative, involving 953,000 specific regions of gut microbial genomic fragments of the 16S ribosomal RNA (rRNA) gene.

In this exploration of the microscopic world, we find that only 43% of cells in the human body are our own, while the majority consist of various microorganisms. These gut microbes are closely associated with human health and diseases, such as diabetes, depression, autism, among others. Researchers, by understanding the types and phylogenetic relationships of microbes in the gut, can further investigate their functions and roles. Similarities in genetic sequences often indicate closer phylogenetic relationships between organisms.

In the game "Borderlands Science," whenever players attempt to align pattern blocks, they are actually assisting scientists in developing a new algorithm that can compare multiple genome sequences at the same time. The reference pattern on the far left of the game is the common point of these nearly a million sequences, generated by computer analysis. Therefore, when players try to match pattern blocks in the game, they are essentially helping to compare the differences between multiple sequences and the median values of standard samples.

This discovery proves that as players enjoy the game, their combination with science can produce unexpected value. Through this interactive approach, not only is the public's engagement and interest in science increased, but it also opens new avenues for scientific research.

In the context of ever-advancing artificial intelligence technology, the role that gamers play in the field of sequence comparison has become particularly important. The reason is that there is no so-called "standard answer" when it comes to sequence comparison, so scientists cannot provide AI with a clear reference like they do with answers to math problems. This task requires finding the optimal solution between two conflicting goals: aligning as many nucleotides as possible while reducing the number of gaps generated. This is similar to the "spacer blocks" in the game.

Let's take the comparison of two sequences, TTCATAGC and CTTGACTG, as an example to illustrate the problem. If we strictly prohibit different nucleotides from lining up together, the final comparison result will contain many gaps. However, such an arrangement provides little information and hardly lets us understand the relationship between the two sequences. Conversely, if we refuse to accept any gaps at all, we ultimately cannot achieve the purpose of comparison.

Let's use our brains and human creativity to seek a compromise solution that displays the similarity between sequences while minimizing the use of "spacer blocks." Such results seem more reasonable, as they consider both resemblance and try to minimize gaps. This is certainly not the only possible method; everyone can try to compare these two sequences using their own method and find the best solution. This reflects the different ways in which humans and computers solve these problems.

In fact, the design of "Borderlands Science" is not to pursue the ultimate in sequence alignment, but to understand the patterns that humans might adopt when dealing with such problems and which ones are preferred by human players. Scientists designed this game because it is difficult to train computers to balance between alignment and reducing gaps. The "intuition" of human players becomes an extremely valuable tool in this process. When facing multiple sequences, humans can consider it from a more global perspective, while computers usually compare sequences in pairs.

As millions of players provide their solutions, scientists are able to select those relatively optimal solutions that are no worse than other solutions to train more efficient computer algorithms. Since the official launch of "Borderlands Science" on April 7, 2020, gamers have contributed over 75 million sequence comparison solutions, with an average of 43 different solutions per puzzle. As of May 2023, over 4 million game players have completed at least one puzzle (excluding tutorial levels), reaching an astonishing participation rate of 90%.

Experiments have shown that the multiple sequence alignment algorithm obtained using the data from "Borderlands Science" has surpassed other advanced algorithms, such as PASTA, MUSCLE, and MAFFT, and has successfully drawn the highest quality phylogenetic tree of the gut microbiome to date. This not only highlights the great achievement of "Borderlands Science" as a citizen science project but also proves it to be one of the largest and most significant citizen science projects in history.

In 2010, an online casual puzzle game named Phylo was created, targeting players to align promoter sequences of mammalian genes, including human DNA sequences. Despite the relatively small data size of Phylo and the game's scientific team's knowledge of the source of each sequence, which allowed them to evaluate players' solutions, it was an early attempt to provide players with an interactive platform for scientific research.

Since the Phylo interface was quite professional, the participation rate of average players was only 10%, which means the game had limited attractiveness to the general public. In an attempt to attract more players, MMOS Company decided in 2019 to develop a simpler, more fun, and widely accessible game. They simplified the gameplay, shortened the game duration, and combined it with a popular online game. Moreover, to further attract players, they introduced compatibility with game controllers, a reward mechanism, and wrote a story for the game arcade related to a popular character in the game to appeal to fans of that character.

The resulting "Borderlands Science" not only enhanced the gaming experience but also made a significant contribution to scientific research. It has helped scientists reconstruct the evolutionary history of the gut microbiome, greatly advancing the understanding and knowledge in this field. Therefore, if you have ever participated in this game, perhaps you have contributed to the advancement of science.