Problems with the denoising step

[Jessica_Bianca_da_Si (Jéssica Bianca da Silva)]

Hi guys,
I have some doubts about the denoising step of the sequences. I've done all the quality assessment and primer removal, but when I perform the denoising, the percentage of input merged is quite low (<50%). I ran it with different truncation values based on the quality of the bases, and finally, I tried it with truncation values equal to 0, and even then the results were not good. Could someone explain why the percentage of input merged is low? Below is an example of the “stats.tsv” table generated from the denoising files:

sample-id	input	filtered	percentage of input passed filter	denoised	merged	percentage of input merged	non-chimeric	percentage of input
SRR21848608	27085	15771	58.23	13276	8359	30.86	5870	21.67
SRR21848609	32438	18318	56.47	15798	10148	31.28	5623	17.33
SRR21848610	32719	19135	58.48	16859	11046	33.76	5181	15.83
SRR21848611	28798	17303	60.08	15328	10656	37.00	5067	17.59

I'd be grateful if anyone could help with this question.
Jessy

[colinvwood (Colin Wood)]

Hello @Jessica_Bianca_da_Si,

How well merging performs can depend on several factors, the most important of which is usually ensuring that your reads have enough overlap at the merging site. Make sure that your amplicon length and paired read lengths (after truncating) allow at least 12bp of overlap.