search

JappyPing / noise2read

MIT License
2 stars 0 forks source link

Comparison to swarm #1

Open colinbrislawn opened 7 months ago

colinbrislawn commented 7 months ago

Good afternoon!

These two tools seem similar in some ways, and I would appreciate your feedback how how they differ in approach and intended usage.

a rare read is erroneous if it has a neighboring read of high abundance

https://github.com/torognes/swarm

swarm is a single-linkage clustering method,... swarm's novelty is its iterative growth process and the use of sequence abundance values to delineate clusters. swarm properly delineates large clusters (high recall), and can distinguish clusters with as little as two differences between their centers (high precision).

I understand that comparison can be contentious. 🙃

I'm less interested in which of these tools is 'better' and more how they can be used best! Plus, you said 'Feel free to contact me' so I'm taking you up on the offer 📫

Thank you for your time!

JappyPing commented 6 months ago

Hi there,

I am very glad to see your messages here. Yeah, feel free to contact me. I'm sorry for not getting back to you sooner; I just saw this issue.

Honestly, I just know Swarm. Yeah, it seems that noise2read and Swarm may have similar descriptions, but I think they are totally different.

Noise2read was designed to eliminate bias and errors produced during the sequencing process; it can be applied to sequencing data with a length of less than 300 bp. In my understanding, noise2read can correct these biases and errors before amplicon clustering, but I am unsure whether it is proper for amplicon clustering directly. I need time to do research and test.

I am highly interested in Swarm and amplicon clustering. Unfortunately, I have been busy with my thesis recently, but I will take more time on Swarm and amplicon clustering next month or later. I will share more with you later if I have a better understanding.

Best regards, Jappy