aqmm

Document for AQMM (Absolute quantification of metagenome and metatranscriptome)

The AQMM was developed to perform absolute quantification of mulitipile metagenome and it's paralell metatranscriptome. In order to use this algorithm, the experiment should be designed with both metagenome and metatranscriptome data. From initial stage, the molecular experimental data should be recorded to help on estimating the overall DNA or RNA of a unit (ml/gram) of sample. The AQMM enable an absolute quantificaton of genes, hence to mitigate the influence of composition effect in relative quantification methods.

clone the source code to local server

git clone https://github.com/biofuture/aqmm.git  

prepare the meta_data.txt which contains all the molecular experimental data  

The meta_data.txt should (items should be seperate by tab "\t") .

SID	SNAME	DNA_weight	DNA_volume	DNA_extr_eff	RNA_weight	RNA_volume	RNA_extr_eff	mRNA_ratio
1	AS1	24570	0.5	0.282	6135	0.5	0.075	0.03
2	AS2	27300	0.5	0.282	6355	0.5	0.075	0.03

Each line contains information of one sample .

SNAME is the name of the sample, which stored under the input directory .
DNA_weight is the weight of DNA for the extraction (ng) .
DNA_volume is the total volume of sample used for the DNA extraction .
DNA_extr_eff is the extracting efficiency for the sample, which is an emperical value for a DNA extraction KIT for a sample .
RNA_weight is the weight of RNA for the extraction .
RNA_volume is the total volume of sample used for the RNA extraction .
RNA_extr_eff is the extracting efficiency for the sample, which is an emperical value for a RNA extraction KIT for a type of sample .
mRNA ratio is the ratio of mRNA to total RNA, which is a emperical value for different environment

Notice: Before using this algorithm, users need to get the above meta data information as accurately as possible, although the algorithm show robusty to the variance to the extraction efficiency among different samples, users should keep all the condition as consistent as possible in order to get more accurate/reliable quantification and comparativety.

Running the aqmm just by one command from sequence file

By typing ./aqmm -h the help information will show here (Perl version 5 (5.1) need to run the program) .  

perl ./aqmm -a <Input_metagenome_dir> -b <Input_metatranscriptome_dir> -m <Experimental_meta_data> -n <num_threads> -o <ouput_dir>
Author Jiang Xiao-Tao 2020-07
Email  [email protected]

where:

   -a <Input_metagenome_dir> including all the fq files of your metagenomic samples after quality filtering
   
   -b <Input_metatranscriptome_dir> including all the paralell metatranscriptomic fq files with metagenomic samples after . remvoing the rRNA with tools like SortMeRNA.
   
   -m meta_data.txt including all the experimental extraction information and the extraction efficiency information of your sample type
   
   -o <ouput_dir> will include all the output results for the quantification, including the esimated cell numbers of metagenomic data of each sample and the estimated sequenced cell numbers your metatranscriptomic data included. With the quantification results obtained, the abundance of gene/species could be futher normalized to per cell/volume/gram etc. al.   

Example to run

There is a small demo to show how to use AQMM to generate absolute quantification of your data under the example directory.

Enter the example directory and run the following command

../dist/aqmm  -a AS_metagenome -b AS_metatranscriptome -m meta_data.txt -n 12 -o testout

After runnning, check the results

cat dna_cell_info.txt .   

SID	SNAME	library_size	lib_cell	cell_per_ml	A_ratio	T_ratio	C_ratio	G_ratio
1	AS2	50000000	8.56742513314301	3.23275242169008e+19	0.22371768	0.22220922	0.27649788	0.27757522
2	AS1	50000000	9.93418747046929	3.37319232480681e+19	0.21272236	0.21128754	0.28672574	0.28926436

cat rna_cell_info.txt .

SID	SNAME	library_size	lib_cell	cell_per_ml	A_ratio	T_ratio	C_ratio	G_ratio
1	AS2	50000000	8.56742513314301	3.23275242169008e+19	0.22371768	0.22220922	0.27649788	0.27757522
2	AS1	50000000	9.93418747046929	3.37319232480681e+19	0.21272236	0.21128754	0.28672574	0.28926436

These numbers could be used to normalize your genes/species to per cell/volume level in the end!

Compare with RQ methods

A script was developed to process the results for both AQMM and RQ methods like RPKM/TPM/edgeR to identify the differential gene expression between groups::

perl compare_aqmm_rq.pl 
Author: Xiao-Tao JIANG
Date:
Modified : 19-01-2018
Email: [email protected]
perl compare_aqmm_rq.pl -a <meta_data_rna.txt> -b <unique_gene_set> -c <rna_reads_count.matrix> -o <Oprefix> -h
-a input meta data file for all the metatranscriptomic data <required>
-b the gene set fasta file <required>
-c matrix of reads number mapped in RNA samples for each gene  <required>
-d matrix of transcript per copy gene  run AQMM required.
-o output prefix, all output result files name are prefix with this string <required>  
-l the length of reads default strategies PE 150 sequencing, default 150 bps
-s this option is to select the normalization methods, RPKM, TPM, EDGER, AQMM, ALL. default AQMM \
-h print this help information 

Generate Transcript per copy gene matrix

After mapping DNA and RNA to the gene set, the DNA and RNA read count matrixs are obtained. For a general analysis, this gene read cont matrix can be futher normalized by gene length, library size etc,al to get quantifications such as RPKM, TPM, TMM. Then statistical analysis can be down on these matrix to identify the diffential abundant genes. For the AQMM package, we have proposed a quantification TPCG, which can show the activity of each copy of gene, this is not the copy of transcipts, becasue in a microbial system,the gene may have many copiesthe present of big number of transcripts for one gene does not neccessarity mean this gene is very active. To generate this matrix, the DNA and RNA matrix need to be normalized to Per cell level. The number of sequenced cells are obtained by aqmm script.

perl TPCG_matrix.pl <DNA_PerCell> <RNA_PerCell> <OUT_TPCG> 
DNA_PerCell  DNA per cell matrix
RNA_PerCell  RNA per cell matrix 
OUT_TPCG     TPCG output