[vintermann]

This is a dataset of bacterial DNA. Any two related bacteria will have long strings of the same letters. But it won't be neatly aligned, so the line breaks will mess up pattern matching.

[bede]

Exactly. The line breaks break the runs of otherwise identical bits in identical sequences. Unless two identical subsequences are exactly in phase with respect to their line breaks, the hashes used for long range matching are different for otherwise identical subsequences.

[amelius]

And the compressor does not think: "how can I make these two sequences align better without wasting a lot of space?"

[ebolyen]

No, because alignment, in the general case, is O(n^2). It is ironically one of the more tractable and well solved problems in bioinformatics.

[tiagod]

The compressor doesn't think about anything. Also, Zstd doesn't have the goal of reaching the highest possible compression ratio. It's more geared toward lowest overhead, high bandwidth compress/decompress.