Metaphlan output

[shashank gupta]

Hi,
i am trying to do MetaPhlAn using Galaxy server, i am using it first time.

As it is mentioned in the tutorial that Metaphlan accepts Fasta sequence which looks like-
>seq1
AAATTGCATTTAGACCAGG
>seq2
ATATATGATAGCGCGATGA
>seq3
CGTGCGCGTATAGTGGGGCTA and so on..

i uploaded fasta file in the GALAXY server, it gets uploaded without any error, then clicked Metaphlan keeping options default and clicked execute, again it got completed without an error.

But when i saw the output it shows something like this

unclassified 100.0

Am i missing something, because when i used the test fasta file available on GALAXY server, it works perfectly.

Best!
Shashank

[Nicola Segata]

Hi Shashank,
 can it be that the size of the dataset you uploaded is too small (i.e. a small number of reads)? In that case MetaPhlAn will likely give you 100 unclassified, because MetaPhlAn is not a read-classifier but an estimator of species abundances and with just few reads the result cannot be robust.

I'd be able to help you more if you tell me the characteristics of the dataset (number of reads, length of the reads, origin of the sequences).

cheers

Nicola

--
You received this message because you are subscribed to the Google Groups "MetaPhlAn-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to metaphlan-users+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[shashank gupta]

Hi,
the file i uploaded have 46,317 sequences.
len min/max/avg 361.0/845.0/655.0


thanks
Shashank

[Nicola Segata]

Hi,
 for MetaPhlAn we are used to use at least half milling shot reads as input. You reads are quite long compared to standard Illumina sequencing, so in MetaPhlAn you should get better results using a local mapping strategy, e.g. "--bt2_ps sensitive-local" rather than the default "  --bt2_ps very-sensitive" (without "-local").

cheers

Nicola

[shashank gupta]

I am using Galaxy server, so how can i use above command ?

[George Weingart]

I believe that in Galaxy, metaphlan2 has the option to select that parm

--
You received this message because you are subscribed to the Google Groups "MetaPhlAn-users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to metaphlan-use...@googlegroups.com.

[e.m...@warwick.ac.uk]

Hi all,

I have the same problem, when I run my metagenomic sequencing samples through MetaPhlAn2 I always get 100% unclassified.

We did whole genome sequencing from environmental samples (paired-end) and got about 4,000,000 reads for each samples. Each sequence is about 250bp long.
I tried to use MetaPhlAn2 with joined reads but also with my paired-end reads, I tried sensitive, very-sensitive and very-sensitive-local but the result is always the same - 100 unclassified.

What am I doing wrong?

Cheers,

Eileen

[Duy Tin Truong]

Hi Eileen,

Can you send me the commands that you used? Besides, can you check if metaphlan2 actually received input through the pipe, or if the input file path exists?

Thanks,

Tin

[Duy Tin Truong]

Hi Eileen,

Can you check the output sam file or the bowtie2 output to see if bowtie2 actually works?

Thanks,

Tin

[e.m...@warwick.ac.uk]

Hi Tin,

I used the following command, but also tried it with very-sensitive and with my joined data rather than forward and reverse reads

metaphlan2.py 1_S1_L001_R1_001.fastq,1_S1_L001_R2_001.fastq --mpa_pkl db_v20/mpa_v20_m200.pkl --bowtie2db bowtie2db/mpa --bowtie2out metagenome.bowtie2_sensitive.bz2 --nproc 5 --bt2_ps very-sensitive-local --input_type fastq > MSAP1_sensitive.txt

How can I check if methaphlan2 received input through the pipe?

I attached the bowtie2 output file (ziped). I had a look into it but do not know if that looks alright.

Cheers,

Eileen

[Duy Tin Truong]

When checking your bowtie2 output, I could not find the marker name in the metaphlan2 database. Besides, looking at the command that you used, I think there was something wrong with bowtie2db option as you are running metaphlan2 with metaphlan1 database, can you try again with this:

cat 1_S1_L001_R1_001.fastq 1_S1_L001_R2_001.fastq | metaphlan2.py  --mpa_pkl db_v20/mpa_v20_m200.pkl --bowtie2db bowtie2db/mpa_v20_m200 --bowtie2out metagenome.bowtie2_sensitive.bz2 --nproc 5 --bt2_ps very-sensitive-local --input_type fastq > MSAP1_sensitive.txt

Thanks,

Tin