Sperm based data storage when?
精子存储数据什么时候出？
精子存储数据什么时候出？
>>109436223
Not funny. Ngmi
不好笑。你完了。
Interesting file system your block size is 37.5MB
你家文件系统挺有意思，块大小 37.5MB。
>>109436223
It's been quite a long time since laptops only had 2GB of storage. This shit is fucking retarded.
笔记本电脑只有 2GB 存储的时代已经过去很久了好吗。这玩意儿纯粹是脑残。
>>109436223
That's totally fucking wrong. Human genome has billions of base pairs, and each pair can be in 1 of 4 configurations
这完全扯淡。人类基因组有几十亿个碱基对，每一对都有 4 种可能配置。
>>109436223
15.000 GB of data.
15000GB 的数据。
That's 15 TB. Also the unique data is 37 MB only, much of which is useless clutter, many times redundant with several write errors (like in OPs case).
那就是 15TB。而且唯一数据只有 37MB，大部分还是没用的垃圾，重复很多，还带一堆写入错误（就像楼主这样）。
>>109436525
the unique data should be up to twice of 37.5.
唯一数据应该是 37.5 的两倍才对。
>>109436336
~3.2 billion base pairs
约 32 亿个碱基对。
One base pair per 2 bits => 4 base pairs per byte, or ~0.8GB per human genome.
一个碱基对占 2 位 => 4 个碱基对等于 1 字节，也就是人类基因组大约 0.8GB。
The way compression in genetics works is that they store a 0.8GB "reference human" and then an individual is a diff on the reference, usually around 1MB.
基因数据的压缩方式是这样的：他们存一个 0.8GB 的“参考人类”，然后再拿个人的数据和参考做 diff，通常只有 1MB 左右。
There's some additional complexity (structural variants, insertions/deletions) but this is the bulk of it.
还有额外一些复杂性（比如结构变异、插入/缺失），但大体上就是这么回事。
>>109436525
"Unique data" is a bit of a misnomer, it's diff-based compression.
“唯一数据”这个说法不太准确，本质上是基于 diff 的压缩。
The 0.8GB is extremely high entropy already so you can't really compress it.
那 0.8GB 本身熵就已经很高了，根本压不动。
It's more like, 0.8GB context + diff of 1MB.
更像是 0.8GB 的上下文加上 1MB 的 diff。
The storage medium itself is 0.8GB for a diploid, or for sperm 0.4GB storage per cell, or one ejaculate contains ~40 petabytes of raw data.
存储介质本身，二倍体要 0.8GB，精子的话每个细胞 0.4GB 存储量，一次射精大概含 40PB 的原始数据。
It's also harder to do the "diff" approach on individual sperms due to the shuffling.
而且由于基因重组的打乱，对单个精子做 diff 方式也更难。
i liked songbird based data storage more where you encode the data in a fourier transform or some shit and play it back to a bird that then sings it
我更喜欢基于鸣禽的数据存储，把数据编码成傅里叶变换之类的玩意儿，然后回放给鸟听，鸟再唱出来。
>>109436223
>Sperm based data storage when?
>精子存储数据什么时候出？
Sperm is generated based on your DNA.
精子是根据你的 DNA 生成的。
It'll always just contain variations on the same data.
它永远只是同一份数据的变体而已。
Creating custom DNA strands with data encoded in them is possible, and being done, but the data rates of reading and writing make it not worthwhile at this point.
把自定义DNA链里编码数据是可行的，而且也在做了，但读写速度让这玩意儿目前不值当。
>>109436525
>much of which is useless clutter
>其中很多是没用的垃圾
not all important dna encodes for genes
不是所有重要的DNA都编码基因啊
dna is also used for controlling gene expression, which is poorly understood
DNA还用来控制基因表达，这块我们理解得很差
dna labeled as "junk" is more a indicator of our ignorance and over-confidence understanding rather than it actually being useless
被标成“垃圾”的DNA更多说明我们无知和过度自信，而不是它真没用
even the position and particular sequences of "junk" basepairs has important ramifications on things like molecular stability of dna
连“垃圾”碱基对的位置和特定序列都对DNA分子稳定性之类有重要影响
>me
You don't look like Arnold
你又不像阿诺德
>>109436252
kek
Genetics for life and genetics for computing are very different. Read Adlman's paper on it.
用于生命的遗传学和用于计算的遗传学完全是两码事。去读读Adlman那篇论文吧。
（如果你觉得这篇文章有启发，可以点击这里付费）
