AlphaSquad
commented
9 months ago Hi, thank you for your interest in CAMISIM.
CAMISIM inherently does not care about the genomes' relatedness and since you assign the NCBI ID of 2 to all of them, it thinks they are all Bacteria anyways. Note that you will get a gold standard assembly in any case, as this cannot be turned off in a feasible way.
In terms of abundance: If you use ART as read simulator, the abundance CAMISIM uses is the depth at which a genome is sequenced - this means you would need to normalise by genome length if you wanted the same number of reads. For Nanosim and wgsim there currently is a bug which ignores genome sizes, so you actually would get the same number of reads.
So generally I think your files look correct. I would recommend simulating a small data set with like 100MB in a single sample and check if the output looks like you would expect. If it does not, feel free to come back.
CorvusVaine
Hello, Rather than an issue with code, I would need an advice on data generation.
I would like to generate a sample of some species using only my own genomes. I only need to generate reads and need no assembly, I just want to make sure I will have the good proportions. For each specie, I have xi consensus genomes, themselves obtained from clustering of yi genomes.
Rather than using one genome by specie, I would like, for the sake of diversity in data generation, to use all the xi * yi genomes linked to the specie.
For example, Iet's say I have 2 species : E. coli and V. parvula.
I have two clusters of genomes for E. coli : E.coli_A and E.coli_B. Cluster A contains 3 genome and B contains 2 genomes, so I have, as a whole, E.coli_A_1, E.coli_A_2, E.coli_A_3, E.coli_B_1 and E.coli_B_2.
V. parvula has one cluster of 4 genomes : V. parvula_A_1, V. parvula_A_2, V. parvula_A_3, V. parvula_A_4.
I want my species to be equally represented, and each genome to be equally represented in the specie. My abundance would therefore be of ½ 1/5 = 1/10 for each E.coli and ½ ¼ = ⅛ for each V.parvula.
How would you recommand designing my abundance file and metadata ?
Can I just consider each genome to be on the same level and have metadata like this :
Genome_ID. OTU. NCBI_ID. novelty_category E.coli_A_1. 1. 2. novel_species E.coli_A_2. 2. 2. novel_species E.coli_A_3. 3 2. novel_species E.coli_B_1. 4 2. novel_species E.coli_B_2. 5. 2. novel_species V.parvula_A_1 6 2. novel_species V.parvula_A_2 7 2. novel_species V.parvula_A_3 8. 2. novel_species V.parvula_A_4 9. 2. novel_species
and abundance like this :
E.coli_A_1. 0.1 E.coli_A_2. 0.1 E.coli_A_3 0.1 E.coli_B_1 0.1 E.coli_B_2 0.1 V.parvula_A_1. 0.125 V.parvula_A_2. 0.125 V.parvula_A_3. 0.125 V.parvula_A_4. 0.125
Or do I have to notify CAMISIM when they are from the same specie ? I figured this repartition would make sure each genome is treated independently to the proportion I hope to get, but I just want to make sure I am not mistaken.
(For the sake of abundance, I was notified that if I wanted each genome to give the exact same number of reads, I should normalize its abundance by its length, as longer genomes will produce more reads, is that so ?)
Thank you for your quick answer.