Need to adapt kmer size values ?

[NJeanray]

Hello Trinity Team,

I'm currently triying to identify alternative transcripts from RNA-Seq Homo Sapiens samples.
I ran Trinity successfully on each of my samples :

Trinity --seqType fq --left sample_R1.fq.gz --right sample_R2.fq.gz --max_memory 20G --CPU 24 --KMER_SIZE 32 --output /trinity_output/

I performed a check of my assembly quality doing an alignment of my initial reads againt their corresponding assembly, and unfortunately, this not really good :

Also, I had a look at the generated de Bruijn graphes and it seems that I get a lot of disconnected nodes. So, I guess this is coming from the length of kmers which are too small. I read that Trinity uses by default a size of 25 and cannot exceed 32.

What would you advice ? The range of my reads lenght is about 36-76 nucleotides (short reads).

What would you do ?

Thanks in advance !

[brianjohnhaas]

Hi, Try using the latest version of Trinity and do not try to change the kmer size. I expect you'll get better results. For comparison, try another assembler like SPADES-RNA. best, ~brian

…

On Wed, Dec 15, 2021 at 4:19 AM NJeanray ***@***.***> wrote: Hello Trinity Team, I'm currently triying to identify alternative transcripts from RNA-Seq Homo Sapiens samples. I ran Trinity successfully on each of my samples : Trinity --seqType fq --left sample_R1.fq.gz --right sample_R2.fq.gz --max_memory 20G --CPU 24 --KMER_SIZE 32 --output /trinity_output/ I performed a check of my assembly quality doing an alignment of my initial reads againt their corresponding assembly, and unfortunately, this not really good : [image: image] <https://user-images.githubusercontent.com/4158621/146157650-890cae7a-dc23-4f80-aefd-a7249bf53fa8.png> Also, I had a look at the generated de Bruijn graphes and it seems that I get a lot of disconnected nodes. So, I guess this is coming from the length of kmers which are too small. I read that Trinity uses by default a size of 25 and cannot exceed 32. What would you advice ? The range of my reads lenght is about 36-76 nucleotides (short reads). What would you do ? Thanks in advance ! — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#1105>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABZRKX2YASWK2NNULPNFPTLURBMSRANCNFSM5KDDEV4Q> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

-- -- Brian J. Haas The Broad Institute http://broadinstitute.org/~bhaas <http://broad.mit.edu/~bhaas>

[NJeanray]

Hi Brian,

Excuse-me, I made a mistake writting the command line : I did not set the parameter regarding the kmer size. So I guess Trinity won't give me better results even if I force this parameter to a given value ?

[brianjohnhaas]

I see. That's correct, then. Take a look at your EXN50 and BUSCO stats before discarding it, though. Then, comparing to another assembler would be helpful to see if another might do better for some reason.

…

On Wed, Dec 15, 2021 at 9:03 AM NJeanray ***@***.***> wrote: Hi Brian, Excuse-me, I made a mistake writting the command line : I did not set the parameter regarding the kmer size. So I guess Trinity won't give me better results even if I force this parameter to a given value ? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#1105 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABZRKX6FX5GSZNSRBRTCQCTURCN3TANCNFSM5KDDEV4Q> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

-- -- Brian J. Haas The Broad Institute http://broadinstitute.org/~bhaas <http://broad.mit.edu/~bhaas>