Phylogeny can not be inferred. No enough markers were kept for the samples

Microbial community profiling StrainPhlAn
1

Hi,

I tried to run StrainPhlan on s__Eisenbergiella_tayi.fna, but I first got the error of “too many samples were discarded”. Then, I set “–sample_with_n_markers 0 --marker_in_n_samples 0”, and it gave me this error “[e] Phylogeny can not be inferred. No enough markers were kept for the samples”.

What does it mean? I first did the Wilcoxon test on some species and I got this species with an adjusted p < 0.05, so I tried it with StrainPhlan. If this error occurs, does it mean this species is not good for analysis?

Thank you,
Ai

2

Hi @Ai_Zhang
Thanks for getting in touch. Could you give us few more information to better understand the problem:

  1. How many samples are you considering?
  2. In how many samples E. tayi is present? Is it low abundant in your samples?
  3. Could you share the exact full command you are using?

Thanks

3

Hi Aitor,

I used 63 samples. Only 4 samples present E. tayi. The relative abundances of these 4 samples from MetaPhlan are 1.6, 0.57, 0.007, and 0.004.

Command I used:
strainphlan -s consensus_markers/*.pkl -m clade_markers/clade_markers/s__Eisenbergiella_tayi.fna -o output -c s__Eisenbergiella_tayi --phylophlan_mode fast --nproc 16 --sample_with_n_markers 0 --marker_in_n_samples 0

I checked another species, s__Bacteroides_vulgatus, that I could run strainphlan with default settings without error. For this species, only 5 samples have it. Relative abundances are 6.3, 1.05, 0.38, 0.09, and 0.02.

Thank you,
Ai

4

Hi @Ai_Zhang
Having only 4 samples with E. tayi present is already a challenge when you have such low abundances in your samples (0.007 and 0.004). StrainPhlAn trees need a minimum of 4 samples and it seems that, at least in one (and probably two) or your samples you don’t have enough reads to reconstruct the enough marker sequences (due to its low abundance). The scenario with B. vulgatus looks better, the low abundance samples are ~10 times higher.
Best,
Aitor

5

Hi Aitor,

Thanks for the reply. It’s really helpful!

Thank you,
Ai