Trinity, reverse complement transcripts, and transdecoder
578 views
Skip to first unread message
David R
Jan 24, 2017, 12:11:31 PM1/24/17
to trinityrnaseq-users
Hello,
I'm trying to assemble and annotate a transcriptome for a non model organism (insect).
My biggest concern at this point (and a point of confusion) is that I still have messed something up. In my final data set of ~23000 ORFs, at least a few ORFs are reverse complements to each other. If this were a fungus I'd be more comfortable with this, but for an insect overlapping genes is less common...so I'm now doubting that I did everything correctly.
I would appreciate some quick verification/reassurance that I'm doing this all correctly, as well as advice for what I should do differently.
Here is what I have done thus far:
1) ran Trinity assembler on a large (Illumina) transcriptome data set using default parameters and obtained a ~120,000 transcript "Trinity.fasta" file
2) ran Transdecoder.LongestOrfs to extract longest cds. This resulted in ~48,000 transcripts.
3) tried numerous times to get Transdecoder.Predict to run, but it kept crashing due to some perl module incompatibilities.
4) noticed that the Trinotate pipeline I plan on using didn't actually require/want Transdecoder.Predict output, so I abandoned that and simply went with the longest_orfs from Transdecoder.LongestOrfs (is this ok? How important is it that I get Transdecoder.Predict to run??)
5) while looking over the transcripts, I noticed that my some of them (in the longerst_orf.cds file) were identical; so I ran the dedupe.sh script from bbmap to remove duplicates. This reduced my ORFs down to ~23,000
6) I ran BUSCO on this file and it is 97% complete (!!!) but it still contains a few duplicate transcripts.
My question--before I proceed--is does all this sound reasonable, correct and fair (to the data)?
Any words of advice?
Thank you.
David
Brian Haas
Jan 24, 2017, 9:18:28 PM1/24/17
to David R, trinityrnaseq-users
Hi David,
please see my comments below:
On Tue, Jan 24, 2017 at 12:11 PM, David R <david....@gmail.com> wrote:
Hello,I'm trying to assemble and annotate a transcriptome for a non model organism (insect).My biggest concern at this point (and a point of confusion) is that I still have messed something up. In my final data set of ~23000 ORFs, at least a few ORFs are reverse complements to each other. If this were a fungus I'd be more comfortable with this, but for an insect overlapping genes is less common...so I'm now doubting that I did everything correctly.
If you're running the assembly in strand-specific mode, then it could be that there's some antisense transcription being assembled. If the transcripts are highly expressed, then you might get some 'fake antisense' assembled (the strand-specific methods are about 99% accurate, and that remaining 1% can accumulate sufficient reads to assemble transcripts).
I would appreciate some quick verification/reassurance that I'm doing this all correctly, as well as advice for what I should do differently.Here is what I have done thus far:1) ran Trinity assembler on a large (Illumina) transcriptome data set using default parameters and obtained a ~120,000 transcript "Trinity.fasta" file2) ran Transdecoder.LongestOrfs to extract longest cds. This resulted in ~48,000 transcripts.3) tried numerous times to get Transdecoder.Predict to run, but it kept crashing due to some perl module incompatibilities.
What's the error? It should be straightforward to resolve this.... missing modules can be installed from CPAN pretty easily.
4) noticed that the Trinotate pipeline I plan on using didn't actually require/want Transdecoder.Predict output, so I abandoned that and simply went with the longest_orfs from Transdecoder.LongestOrfs (is this ok? How important is it that I get Transdecoder.Predict to run??)
Trinotate absolutely wants the final TransDecoder predictions. The LongestOrfs will include many false positives.
5) while looking over the transcripts, I noticed that my some of them (in the longerst_orf.cds file) were identical; so I ran the dedupe.sh script from bbmap to remove duplicates. This reduced my ORFs down to ~23,000
Some alt splice variations occur just in the UTR regions, and so the CDS sequences would be identical in this case.
6) I ran BUSCO on this file and it is 97% complete (!!!) but it still contains a few duplicate transcripts.
Duplication rates can be high due to alternatively spliced transcripts being represented. It's not necessarily a bad thing... the duplication stats are more useful when contrasting genome assemblies than transcriptome assemblies.
97% complete is pretty great!! Nicely done!
My question--before I proceed--is does all this sound reasonable, correct and fair (to the data)?
sounds good to me. Let's try to get TransDecoder working for you.
Any words of advice?
Check out our various guides at:
(panel at right)
for more ways of exploring the quality of your assembly.
best of luck!
~brian
--Thank you.David
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsub...@googlegroups.com.
To post to this group, send email to trinityrnaseq-users@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.
Message has been deleted
David R
Jan 25, 2017, 11:48:49 AM1/25/17
to trinityrnaseq-users, david....@gmail.com
Hi and thanks,
Here's the screen dump from the "runMe.sh" (using version 3.0.1)
The error is toward the bottom. I'm not familiar enough with perl to know what needs fixing/installing.
FWIW, I tried installing BerkelyDB6 and adding it to my path but I still get the error.
##############################################
[userD@vmps09 pasa_example]$ bash runMe.sh
-first extracting base frequencies, we'll need them later.
CMD: /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/compute_base_probs.pl pasa_assemblies.fasta 0 > pasa_assemblies.fasta.transdecoder_dir/base_freqs.dat
CMD: touch pasa_assemblies.fasta.transdecoder_dir/base_freqs.dat.ok
- extracting ORFs from transcripts.
-total transcripts to examine: 858
[800/858] = 93.24% done
#################################
### Done preparing long ORFs. ###
##################################
Use file: pasa_assemblies.fasta.transdecoder_dir/longest_orfs.pep for Pfam and/or BlastP searches to enable homology-based coding region identification.
Then, run TransDecoder.Predict for your final coding region predictions.
CMD: /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/get_top_longest_fasta_entries.pl pasa_assemblies.fasta.transdecoder_dir/longest_orfs.cds 5000 > pasa_assemblies.fasta.transdecoder_dir/longest_orfs.cds.top_longest_5000
CMD: /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/bin/cd-hit-est -r 1 -i pasa_assemblies.fasta.transdecoder_dir/longest_orfs.cds.top_longest_5000 -T 1 -c 0.80 -o pasa_assemblies.fasta.transdecoder_dir/longest_orfs.cds.top_longest_5000.nr80 -M 0
================================================================
Program: CD-HIT, V4.6 (+OpenMP), Dec 17 2016, 13:11:13
Command:
/gpfs22/home/userD/Software/TransDecoder-3.0.1/util/bin/cd-hit-est
-r 1 -i
pasa_assemblies.fasta.transdecoder_dir/longest_orfs.cds.top_longest_5000
-T 1 -c 0.80 -o
pasa_assemblies.fasta.transdecoder_dir/longest_orfs.cds.top_longest_5000.nr80
-M 0
Started: Wed Jan 25 10:27:51 2017
================================================================
Output
----------------------------------------------------------------
total seq: 1067
longest and shortest : 4278 and 300
Total letters: 872784
Sequences have been sorted
Approximated minimal memory consumption:
Sequence : 1M
Buffer : 1 X 12M = 12M
Table : 1 X 16M = 16M
Miscellaneous : 4M
Total : 34M
Table limit with the given memory limit:
Max number of representatives: 4000000
Max number of word counting entries: 563639000
comparing sequences from 0 to 1067
.
1067 finished 296 clusters
Apprixmated maximum memory consumption: 37M
writing new database
writing clustering information
program completed !
Total CPU time 6.49
CMD: /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/get_top_longest_fasta_entries.pl pasa_assemblies.fasta.transdecoder_dir/longest_orfs.cds.top_longest_5000.nr80 500 > pasa_assemblies.fasta.transdecoder_dir/longest_orfs.cds.top_500_longest
CMD: touch pasa_assemblies.fasta.transdecoder_dir/longest_orfs.cds.top_500_longest.ok
CMD: /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/seq_n_baseprobs_to_logliklihood_vals.pl pasa_assemblies.fasta.transdecoder_dir/longest_orfs.cds.top_500_longest pasa_assemblies.fasta.transdecoder_dir/base_freqs.dat > pasa_assemblies.fasta.transdecoder_dir/hexamer.scores
CMD: touch pasa_assemblies.fasta.transdecoder_dir/hexamer.scores.ok
CMD: /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/score_CDS_liklihood_all_6_frames.pl pasa_assemblies.fasta.transdecoder_dir/longest_orfs.cds pasa_assemblies.fasta.transdecoder_dir/hexamer.scores > pasa_assemblies.fasta.transdecoder_dir/longest_orfs.cds.scores
CMD: touch pasa_assemblies.fasta.transdecoder_dir/longest_orfs.cds.scores.ok
#####################
Counts of kept entries according to attributes:
FRAMESCORE 483
FRAMESCORE|LONGORF 329
LONGORF 2
########################
CMD: /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/index_gff3_files_by_isoform.pl pasa_assemblies.fasta.transdecoder_dir/longest_orfs.gff3
Can't locate DB_File.pm in @INC (@INC contains: /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/../PerlLib /usr/local/perl/latest/x86_64/sysgcc/nonet/lib/5.14.2/x86_64-linux-thread-multi /usr/local/perl/latest/x86_64/sysgcc/nonet/lib/5.14.2 /usr/local/perl/latest/x86_64/sysgcc/nonet/lib /home/userJ/perl5/bin/lib/perl5/ /home/userJ/perl5/lib/x86_64-linux-thread-multi /home/userJ/perl5/lib /home/userJ/perl5/lib/perl5/x86_64-linux-thread-multi /home/userJ/perl5/lib/perl5 /home/userA/perl5/lib/perl5/x86_64-linux /home/userA/perl5/lib/perl5/x86_64-linux-thread-multi /home/userA/perl5/lib/perl5 /home/userJ/perl5/biomart-perl/lib /usr/local/perl/5.14.2/x86_64/sysgcc/nonet/lib/site_perl/5.14.2/x86_64-linux-thread-multi /usr/local/perl/5.14.2/x86_64/sysgcc/nonet/lib/site_perl/5.14.2 /usr/local/perl/5.14.2/x86_64/sysgcc/nonet/lib/5.14.2/x86_64-linux-thread-multi /usr/local/perl/5.14.2/x86_64/sysgcc/nonet/lib/5.14.2 .) at /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/../PerlLib/TiedHash.pm line 6.
BEGIN failed--compilation aborted at /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/../PerlLib/TiedHash.pm line 6.
Compilation failed in require at (eval 6) line 2.
...propagated at /usr/local/perl/latest/x86_64/sysgcc/nonet/lib/5.14.2/base.pm line 94.
BEGIN failed--compilation aborted at /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/../PerlLib/Gene_obj_indexer.pm line 6.
Compilation failed in require at /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/index_gff3_files_by_isoform.pl line 8.
BEGIN failed--compilation aborted at /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/index_gff3_files_by_isoform.pl line 8.
Error, cmd: /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/index_gff3_files_by_isoform.pl pasa_assemblies.fasta.transdecoder_dir/longest_orfs.gff3 died with ret 512 at ../../TransDecoder.Predict line 379.
Can't locate DB_File.pm in @INC (@INC contains: /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/../PerlLib /usr/local/perl/latest/x86_64/sysgcc/nonet/lib/5.14.2/x86_64-linux-thread-multi /usr/local/perl/latest/x86_64/sysgcc/nonet/lib/5.14.2 /usr/local/perl/latest/x86_64/sysgcc/nonet/lib /home/userJ/perl5/bin/lib/perl5/ /home/userJ/perl5/lib/x86_64-linux-thread-multi /home/userJ/perl5/lib /home/userJ/perl5/lib/perl5/x86_64-linux-thread-multi /home/userJ/perl5/lib/perl5 /home/userA/perl5/lib/perl5/x86_64-linux /home/userA/perl5/lib/perl5/x86_64-linux-thread-multi /home/userA/perl5/lib/perl5 /home/userJ/perl5/biomart-perl/lib /usr/local/perl/5.14.2/x86_64/sysgcc/nonet/lib/site_perl/5.14.2/x86_64-linux-thread-multi /usr/local/perl/5.14.2/x86_64/sysgcc/nonet/lib/site_perl/5.14.2 /usr/local/perl/5.14.2/x86_64/sysgcc/nonet/lib/5.14.2/x86_64-linux-thread-multi /usr/local/perl/5.14.2/x86_64/sysgcc/nonet/lib/5.14.2 .) at /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/../PerlLib/TiedHash.pm line 6.
BEGIN failed--compilation aborted at /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/../PerlLib/TiedHash.pm line 6.
Compilation failed in require at (eval 6) line 2.
...propagated at /usr/local/perl/latest/x86_64/sysgcc/nonet/lib/5.14.2/base.pm line 94.
BEGIN failed--compilation aborted at /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/../PerlLib/Gene_obj_indexer.pm line 6.
Compilation failed in require at /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/../PerlLib/GFF3_utils.pm line 12.
BEGIN failed--compilation aborted at /gpfs22/home/userD/Software/TransDecoder-3.0.1/util/../PerlLib/GFF3_utils.pm line 12.
Compilation failed in require at ../../util/cdna_alignment_orf_to_genome_orf.pl line 10.
BEGIN failed--compilation aborted at ../../util/cdna_alignment_orf_to_genome_orf.pl line 10.
Done. See pasa_assemblies.fasta.transdecoder.\*
[userD@vmps09 pasa_example]$
Brian Haas
Jan 25, 2017, 12:11:31 PM1/25/17
to David R, trinityrnaseq-users
Hi David,
This is what you need:
If you're able to do direct installations, you can try:
perl -MCPAN -e shell
install DB_File
and see if that works.
best,
~brian
--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsub...@googlegroups.com.
To post to this group, send email to trinityrnaseq-users@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.
David R
Jan 25, 2017, 2:09:05 PM1/25/17
to trinityrnaseq-users, david....@gmail.com
Thanks.Yeah, I've been battling with that so just wondered if it was something obvious. I've already manually installed it (and the make test was successful). The correct path is in @INC too...but it just keeps failing as I pasted above...
I'll keep working on it.
To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.
Brian Haas
Jan 25, 2017, 4:28:33 PM1/25/17
to David R, trinityrnaseq-users
same error?
You might need this to be installed to:
but if the error has to do with not finding the perl lib, that would be weird if it's in @INC.
Setting up the PERL5LIB env var to include the path to the library installation directory will set up the perl @INC to include it at runtime.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsubscribe...@googlegroups.com.
To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.
--
--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsub...@googlegroups.com.
To post to this group, send email to trinityrnaseq-users@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.
David R
Jan 27, 2017, 4:54:17 PM1/27/17
to trinityrnaseq-users, david....@gmail.com
Hi Brian,
I (finally) got it working Installing berkeley db locally (on our cluster) was challenging, and the config file for DB_File then needed to be edited or else DB_File would not install correctly (but w/o throwing an error!). So I've finally been able to run Transdecoder.Predict on my data
So here's my question now :)
I collapsed my raw Trinity output (Trinity.fasta) in two ways 1) the "correct way" (Trinity.fasta>Transdecoder.LongestORF and Trinity.fasta>Transdecoder.Predict to output a the file Trinity.fasta.transdecoder.pep) and 2) the way I outlined above which did not use Transdecoder.Predict but just used dedupe.pl to remove duplicated transcripts
The surprising prat is that my output using the non-Transdecoder.Predict step looks slightly better! Testing with BUSCO, the set that I got using Transdecoder.Predict has 77 duplicated genes while my method shows 47 duplicates. Moreover, my set contained three (!!!) additional BUSCO than the set generated by Transdecoder.Predict.
Does this sound strange to you?
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsub...@googlegroups.com.
To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.
--
--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsub...@googlegroups.com.
To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.
Brian Haas
Jan 28, 2017, 10:26:42 AM1/28/17
to David R, trinityrnaseq-users
Hi David,
The duplicates will show up in the Trinity assembly largely due to alternative splicing, and will also include those cases where alt-splice variations are restricted to the UTR regions of the transcripts (ie. cds sequences will be identical, but cdna sequences are different). Of course, deduplication will remove those, but the importance of running such a step really depends on what your downstream application is. We don't do this by default because of how the data model works in Trinotate (single 'gene' -> 'multiple isoforms', and each isoform has it's own coding region and attributes)
gene ----> isoformA -----> proteinA
\
------>isoformB -----> proteinB
and while isoformA != isoformB, it's sometimes the case that proteinA==proteinB
It's also entirely possible that Transdecoder.predict might not report all true coding regions. There are a number of reasons for this - but based on the small numbers that you're working with, it could be that the training of the Markov model was underpowered by the small number of sequences.
TransDecoder does allow you to include blast results and pfam results based on the long-orf models so as to be sure that you do retain those entries that have clear sequence homology (or conserved domains) regardless of how the Markov model scoring goes. It's a 'safety' mechanism so you don't lose ORFs that you might care about.
I'm glad to hear things are working for you now! It's too bad that the DB_file thing was such a pain to deal with. I'll aim to remove that from a future version to make it more straightforward.
best,
~brian
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsubscribe...@googlegroups.com.
To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.
--
--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsubscribe...@googlegroups.com.
To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsub...@googlegroups.com.
To post to this group, send email to trinityrnaseq-users@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.
David R
Jan 28, 2017, 11:02:58 AM1/28/17
to trinityrnaseq-users, david....@gmail.com
Thanks Brian. When you say "based on the small numbers that you're working with" what do you mean exactly? I began with ~600million long reads that were assembled into ~127,000 "transcripts" (Trinity.fasta) and then paired down to ~47,000 longest ORFs (>100aa) which were further predicted to represent ~28,000 gene isoforms.
To be clear, I'm mostly suspicious that more than 1/3 of the BUSCOs I'm testing are showing up as multiple isoforms. That just (intuitively) seems like a lot given the nature of the particular subset of genes BUSCO is testing
I'm going to do some further evaluation to see what is/is not left For calculating differential expression I feel that I want to remove as many isoforms as possible (rather than dilute reads that may be mapping to assembly artifacts).
Thanks again.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsub...@googlegroups.com.
To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.
--
--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsub...@googlegroups.com.
To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsub...@googlegroups.com.
To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.
Brian Haas
Jan 28, 2017, 11:59:02 AM1/28/17
to David R, trinityrnaseq-users
Gotcha. Forget my comments about small numbers.... the BUSCO set it's tested against is small and that's what caught my attention.
Sure - include BLAST and Pfam results as part of the TransDecoder.predict step and you'll capture those proteins. In either case, if you run the Trinotate protocol, you'll still capture functionally relevant homologies via the blastx search, and so capturing all proper ORFs isn't essential as far as capturing the basic biological representation.... but of course still good to get it right. ;-)
Also, I wouldn't worry about filtering transcripts. Just do the DE analysis at the 'gene' level. You can use the Trinity 'gene' groupings or run a tool like CORSET to define them.
I'll have some new tools integrated very soon (in the next week, hopefully) for studying Trinity alternative splicing.
best,
~b
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsubscribe...@googlegroups.com.
To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.
--
--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsubscribe...@googlegroups.com.
To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsubscribe...@googlegroups.com.
To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-users+unsub...@googlegroups.com.
To post to this group, send email to trinityrnaseq-users@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.
Reply all
Reply to author
Forward
0 new messages