Hi, thanks for the video! It's really helpful. Is there a tutorial on downstream analysis of picrust2 results (eg: plotting heatmaps, etc.). That would be really helpful, thank you!
Suet, unfortunately, I have not put together downstream analytical tutorials in the form of videos. I will consider doing that in the future, though! If you have any specific questions about post-hoc analyses feel free to email me, I would be happy to help. mcashay@gmail.com
Hi Mara, thanks for the very informative video. I have managed to work through the Picrust2 pipeline but am now stumped as the best way to visualize the output. Would it be possible to email you for advice on how to go about doing this?
Thank you Mara, for the tutorial. I have a question about the repeated data. Do we need to take care of repeated samples for the PICRUST2 prediction method? What are the methods available for downstream analysis after the prediction, that take care of repeated measure data?
Hi Binu, I would think that this is dependent on the question that you want to answer. If you want to see how functions change over time, I would suggest that you run PICRUSt2 on your samples individually and then incorporate a repeated measure in your analysis. DESeq2 and corncob are supposed to have the capacity to incorporate repeated measures in the model designs. I would have to refer you to their help/user boards because that's where you would find that information. LMK if you have any other questions!
Thank you :) You will need to find your 'representative sequence' file. That will serve as your rep-seps.fna file. Let me know if you have problems identifying the file but it should be an output from the OTU clustering step - you would likely have a better rep-seqs file to use that is the output from removing chimeric sequences.
Hi Mara, thank you very much for your video, its really helpful! I was wondering if you did rarefy your OTU table before incorporing it into picrust2 or you think rarefying the table "path_abun_unstrat_descript.tsv" after is a better way? And a second question, would you use the rarefied table to do diversity analyses and the unrarefied table to proceed with DESeq2 for example? Thank you very much in advance for your help.
Hi! These are really great questions. I did not rarefy my ASV table before I started the PICRUSt2 pipeline and I also did not rarefy my Pathway Abundance while, either. For me, since I was using DESEq2 for my downstream analyses, I did not want to rarefy the table. I did not perform any diversity analyses that would be dependent on equal sample sizes. I would suggest that you find a paper close to your field of study that implemented PICRUSt and see what they chose to do as far as rarefaction is concerned. Many researchers would argue that rarefaction should never be performed because you are removing some of your data to fit older algorithms that cannot handle unequal sample sizes and there are more recent algorithms that can be implemented (DESEq2, divnet, Corncob, etc). Needless to say, it is complicated and there isn't necessarily a perfect answer to the question of whether or when to rarefy. Good luck!
I hope you are a little less confused after watching this... If you still are, is there a specific prat of the metagenome_predicted_contribution.tsv file that you are most confused by?
@@maracloutier4277 Emm... actually I'am trying to reproduce some post-analysis (like what was done in the PICRUSt2 preprint Figure5) with this file using my own data. I wonder is there are some other analysis tools (like ALDEx2, STAMP) for exploring the information this file provided? thanks🙌
@@liutong369 Yes, there are other tools that can be used. I decided to use DESeq2 because I found a paper that used it. I'm still uncertain what statistical tools are "best" to use for PICRUSt output.
Hi Mara, i am wondering what downstream analysis did you use for your picrust2 result (e.g., enrichment, differential abundance)? Did you use DeSeq2? Thank you! :)
Angela, I use DeSeq2, though corncob is a newer differential abundance analysis that is promising. If I have a categorial variable, I would run corncob, and then incorporate the differentially abundant ECs/KOs into some sort of pathway analysis. KEGG pathways are really useful!
@@maracloutier4277 thank you for your response! :) before you run the data to deseq, did you first convert this to phyloseq data then DeSeq2? :) If it is okay with you, can I send you an email for my elaborate question? Thank you very much. You are very helpful! :)
HI Mara, thank you for the video. I am having problem with exporting data from dada2 (or phyloseq) to biom and fna format. Is there any way I could contact you directly with my questions?
this video is really amazing. really needed some visuals on this. I would like to ask a question though, what file do we start with to get the .fna file. i got the .biom file already. where do i start to get the .fna file. like now i have the fastq files for the 5 samples and their metadata. Thank you very much
Thank you :) My fna file (seqtab) is a file that is generated during the DADA2 pipeline. You should also have a similar file from QIIME2 (rep-seqs.qzv).
@@maracloutier4277 I thought I could go back to the fastq.gz files. And follow a pipeline to the fna and biom file all over. I think I got stuck somewhere in between.
@@maracloutier4277 i did the qiime2 all over and got the files. but now in picrust i have another error. Error running this command: hmmalign --trim --dna --mapali /home/enuh/Desktop/picrust2-2.3.0-b/picrust2/default_files/prokaryotic/pro_ref/pro_ref.fna.gz --informat FASTA -o intermediate/place_seqs/query_align.stockholm /home/enuh/Desktop/picrust2-2.3.0-b/picrust2/default_files/prokaryotic/pro_ref/pro_ref.hmm ../dna-sequences.fasta Standard error of the above failed command: Error: Failed to open sequence file ../dna-sequences.fasta for reading thats the error i get now
Hi Mara, thanks for your great explanation about PICRUST2 pipeline, but I have some problem (unknown parsing error) with visualizing the PICRUST2 using these files: path_abun_unstrat_descrip.tsv, metadata.tsv. Do you know what the problem is?
4 years later this video has been a life-saver, thank you queen
Hi, thanks for the video! It's really helpful. Is there a tutorial on downstream analysis of picrust2 results (eg: plotting heatmaps, etc.). That would be really helpful, thank you!
Suet, unfortunately, I have not put together downstream analytical tutorials in the form of videos. I will consider doing that in the future, though! If you have any specific questions about post-hoc analyses feel free to email me, I would be happy to help. mcashay@gmail.com
Hi, Mara, is there a visualization follow-up steps to this pipeline? Thanks!
Hi Mara, thanks for the very informative video. I have managed to work through the Picrust2 pipeline but am now stumped as the best way to visualize the output. Would it be possible to email you for advice on how to go about doing this?
Thank you Mara, for the tutorial. I have a question about the repeated data. Do we need to take care of repeated samples for the PICRUST2 prediction method? What are the methods available for downstream analysis after the prediction, that take care of repeated measure data?
Hi Binu, I would think that this is dependent on the question that you want to answer. If you want to see how functions change over time, I would suggest that you run PICRUSt2 on your samples individually and then incorporate a repeated measure in your analysis. DESeq2 and corncob are supposed to have the capacity to incorporate repeated measures in the model designs. I would have to refer you to their help/user boards because that's where you would find that information. LMK if you have any other questions!
Thanks, it's a great video, i would like to ask a question though, how can I do these steps with an OTU table and not with an ASV table?
Thank you :) You will need to find your 'representative sequence' file. That will serve as your rep-seps.fna file. Let me know if you have problems identifying the file but it should be an output from the OTU clustering step - you would likely have a better rep-seqs file to use that is the output from removing chimeric sequences.
Hi Mara, thank you very much for your video, its really helpful! I was wondering if you did rarefy your OTU table before incorporing it into picrust2 or you think rarefying the table "path_abun_unstrat_descript.tsv" after is a better way? And a second question, would you use the rarefied table to do diversity analyses and the unrarefied table to proceed with DESeq2 for example? Thank you very much in advance for your help.
Hi! These are really great questions. I did not rarefy my ASV table before I started the PICRUSt2 pipeline and I also did not rarefy my Pathway Abundance while, either. For me, since I was using DESEq2 for my downstream analyses, I did not want to rarefy the table. I did not perform any diversity analyses that would be dependent on equal sample sizes. I would suggest that you find a paper close to your field of study that implemented PICRUSt and see what they chose to do as far as rarefaction is concerned. Many researchers would argue that rarefaction should never be performed because you are removing some of your data to fit older algorithms that cannot handle unequal sample sizes and there are more recent algorithms that can be implemented (DESEq2, divnet, Corncob, etc). Needless to say, it is complicated and there isn't necessarily a perfect answer to the question of whether or when to rarefy. Good luck!
@@maracloutier4277 Thank you so much Mara for your inputs! Really helpful. Wishing you the best of luck with your projects.
Very nice. I am also a PhD student
awsome job, the items in metagenome_predicted_contributtion.tsv do confuse me quite a while
I hope you are a little less confused after watching this... If you still are, is there a specific prat of the metagenome_predicted_contribution.tsv file that you are most confused by?
@@maracloutier4277 Emm... actually I'am trying to reproduce some post-analysis (like what was done in the PICRUSt2 preprint Figure5) with this file using my own data. I wonder is there are some other analysis tools (like ALDEx2, STAMP) for exploring the information this file provided? thanks🙌
@@liutong369 Yes, there are other tools that can be used. I decided to use DESeq2 because I found a paper that used it. I'm still uncertain what statistical tools are "best" to use for PICRUSt output.
Hi Mara, i am wondering what downstream analysis did you use for your picrust2 result (e.g., enrichment, differential abundance)? Did you use DeSeq2? Thank you! :)
Angela, I use DeSeq2, though corncob is a newer differential abundance analysis that is promising. If I have a categorial variable, I would run corncob, and then incorporate the differentially abundant ECs/KOs into some sort of pathway analysis. KEGG pathways are really useful!
@@maracloutier4277 thank you for your response! :) before you run the data to deseq, did you first convert this to phyloseq data then DeSeq2? :) If it is okay with you, can I send you an email for my elaborate question? Thank you very much. You are very helpful! :)
@@chakadoodlesness Yes, I love phyloseq objects! Yes, feel free to email me at mcashay@gmail.com
@@maracloutier4277 Thank you! You’re very helpful! :) I sent you an email for some inquiry :)
HI Mara, thank you for the video. I am having problem with exporting data from dada2 (or phyloseq) to biom and fna format. Is there any way I could contact you directly with my questions?
Hi Daniel, sure thing. I had a lot of problems with the exports, too. Email me mcashay at gmail
@@maracloutier4277 Hi Mara, thank you for the tutorial. I am having problems exporting data from dada2 too. can you help me?
this video is really amazing. really needed some visuals on this. I would like to ask a question though, what file do we start with to get the .fna file. i got the .biom file already. where do i start to get the .fna file. like now i have the fastq files for the 5 samples and their metadata. Thank you very much
Thank you :) My fna file (seqtab) is a file that is generated during the DADA2 pipeline. You should also have a similar file from QIIME2 (rep-seqs.qzv).
@@maracloutier4277 . Thank you very much.
@@enuhblaise7502 Let me know if you cannot find the file and I can try to help (muc345@psu.edu)
@@maracloutier4277 I thought I could go back to the fastq.gz files. And follow a pipeline to the fna and biom file all over. I think I got stuck somewhere in between.
@@maracloutier4277 i did the qiime2 all over and got the files. but now in picrust i have another error.
Error running this command:
hmmalign --trim --dna --mapali /home/enuh/Desktop/picrust2-2.3.0-b/picrust2/default_files/prokaryotic/pro_ref/pro_ref.fna.gz --informat FASTA -o intermediate/place_seqs/query_align.stockholm /home/enuh/Desktop/picrust2-2.3.0-b/picrust2/default_files/prokaryotic/pro_ref/pro_ref.hmm ../dna-sequences.fasta
Standard error of the above failed command:
Error: Failed to open sequence file ../dna-sequences.fasta for reading
thats the error i get now
Hi Mara, thanks for your great explanation about PICRUST2 pipeline, but I have some problem (unknown parsing error) with visualizing the PICRUST2 using these files: path_abun_unstrat_descrip.tsv, metadata.tsv. Do you know what the problem is?
mam great explanation but where is the github link for plotting in R
dawg