News for hackers. Stuff that matters.
Re:Exploring DNA with Deep Learning by mschuster91 on Saturday August 17, 2019 @03:57AM (#20722807)
The non-differing parts are worthless as you are looking for differences. For example, take a group of five people with highly sensitive hearing and a hundred people with normal hearing. You don't care what the sensitives have together with the normals, you care about if you can find a mutation present in all sensitives but in none of the normals. Therefore it makes sense to only look at the differences and throw away the equal parts, you just waste computational resources.
Re:Exploring DNA with Deep Learning by longtom on Saturday August 17, 2019 @06:21AM (#20723214)
But isn't genetic code transcribed sequentially and in the context of previously decoded structures? So it is similar to reading a text that references parts elsewhere in the text. If you only keep the changes between multiple subtly mutated texts, then it becomes an enigma what these changes actually mean and, what's worse, the changes are now in context of nearby changes which may make create completely different meanings.