Metatranscriptome without paired metagenome

[gaofeng-ni]

Hi,

Thanks for the cool program. A few question regarding running Humann for metatranscriptomes without paired metagenome.

As a first trial, I searched against UniRef90. The resulted genefamily.tsv don't have annotation for each gene:

# Gene Family   AF1_Abundance-RPKs      AF2_Abundance-RPKs      AF3_Abundance-RPKs      BE1_Abundance-RPKs      BE2_Abundance-RPKs      BE3_Abundance-RPKs
UNMAPPED        0.6573  0.642728        0.602355        0.752705        0.694749        0.665102
UniRef90_A0A009DWL0     0       1.51238e-07     0       0       0       0
UniRef90_A0A009DWL0|unclassified        0       1.51238e-07     0       0       0       0
UniRef90_A0A009EC87     0       4.07372e-06     2.97242e-06     0       0       0
UniRef90_A0A009EC87|unclassified        0       4.07372e-06     2.97242e-06     0       0       0
UniRef90_A0A009ECP6     0       0       2.60899e-06     0       0       0
UniRef90_A0A009ECP6|unclassified        0       0       2.60899e-06     0       0       0
UniRef90_A0A009EGZ3     2.77973e-06     1.90415e-06     0       0       0       0
UniRef90_A0A009EGZ3|unclassified        2.77973e-06     1.90415e-06     0       0       0       0

1. Since in the documentation includes functional annotation for each gene family, I'm wondering if I missed anything or I need to run annotation separately.
2. I'm working with poorly characterized microbial communities, since my mapping rate is low, it's good to try UniRef50?
3. I have 2 conditions (in triplicates), and I wish to carry out differential expression analysis downstream, what would be the recommended input? Will a normalized and joint table with humann_renorm_table --units cpm or joint tables without normalization and follow with something like DESeq2?

Many thanks!

My codes are:

humann \
--input ${READ_DIR}/$SAMPLE.fastq.gz \
--output $OUTPUT_DIR \
--threads 24 \
--search-mode uniref90 \
--nucleotide-identity-threshold 50 \
--translated-identity-threshold 30 \
--translated-query-coverage-threshold 50 \
--nucleotide-query-coverage-threshold 70 \
--metaphlan-options "--read_min_len 29"

conda deactivate