Question: Machine learning Algorithms on DNA sequences
0
gravatar for husainbioinfos
9 months ago by
pune
husainbioinfos0 wrote:

i am working machine learning project,i have 4 different sequences of human from different regions and want to develop machine learning model

dna ml R • 360 views
ADD COMMENTlink written 9 months ago by husainbioinfos0
1

That's great and all, but it isn't a question.

A model for what?

What sequences?

This question is unanswerable.

ADD REPLYlink written 9 months ago by Joe18k

The sequences are not normal patient sequences ,so there must be common mutation(due to disease ) between all those sequences, on that basis i want to develop model . I am confused that from where i should start ..Any example or something else will really help full.

ADD REPLYlink written 9 months ago by husainbioinfos0

Right, so if there's a mutation - why do you need to arbitrarily apply ML to this problem? Variant calling pipelines are well established.

You also haven't told us what the data is still. Have you got reads? Whole genomes? What state is the data in? Has it been QC'd?

ADD REPLYlink written 9 months ago by Joe18k

whole Genome of patients..the data is in RAW form(Simple FASTA format). ML model for predict same type of sequence using RAW sequences.

ADD REPLYlink written 9 months ago by husainbioinfos0

What will the model predict that non-ML methods do not already do?

ADD REPLYlink written 9 months ago by _r_am32k

I don't see how ML fits with this question?

You have a dataset of sequences with mutation X. You receive new data, and need to check if it has mutation X or not. Why guess this with ML, when you can literally just look at the base-pair position in the new data and see if it has mutation X or not?

ADD REPLYlink written 9 months ago by Joe18k

With respect, if you do not even know where to start and if this even will be helpful then maybe are more well-defined project might make sense and most importantly => an experienced supervisor is required.

ADD REPLYlink written 9 months ago by ATpoint44k
1

What do you want the machine to learn? A model is supposed to learn to predict something tangible based on other tangible input. You only have sequence data, what do you wish to predict from that?

ADD REPLYlink written 9 months ago by _r_am32k

The sequences are not normal patient sequences ,so there must be common mutation(due to disease ) between all those sequences, on that basis i want to develop model . I am confused that from where i should start ..Any example or something else will really help full.

ADD REPLYlink written 9 months ago by husainbioinfos0

Please don't copy/paste the same content. You don't need to reply to each comment if you don't have anything specific to say for that comment.

ADD REPLYlink written 9 months ago by GenoMax94k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1411 users visited in the last hour
_