# HUMANN 3.0 function downstream analysis **URL:** \ **Category:** HUMAnN\ **Created:** [March 14, 2022, 9:42pm UTC](https://forum.biobakery.org/t/humann-3-0-function-downstream-analysis/3311 "2022-03-14T21:42:08Z")\ **Posts on this page:** 4\ **Page:** 1
**Author:** ![sarahi](https://yyz1.discourse-cdn.com/flex027/user_avatar/forum.biobakery.org/sarahi/32/1316_2.png) [@sarahi](https://forum.biobakery.org/u/sarahi)\ **Post date:** [March 14, 2022, 9:42pm UTC](https://forum.biobakery.org/t/humann-3-0-function-downstream-analysis/3311/1 "2022-03-14T21:42:08Z")
Hi, I’m sorry for asking such a rudimentary question, but I’m really new to this. Basically, I’ve run my samples through HUMANN 3.0, and I’d like to do further analysis on the functional data; specifically the pathway abundances, and I’m completely lost on what to do next and what tools to use. Maybe I’m overthinking it but reading the descriptions of all the tools have confused me more. I have joined all my samples together in one table, and I tried to put it through MaAsLin 2 and I get this output: ```auto Error in FUN(newX[, i], ...) : invalid 'type' (character) of argument In addition: Warning message: In vegan::decostand(features_norm, method = "total", MARGIN = 1, : input data contains negative entries: result may be non-sense ``` I made sure that the sample columns lined up for both the function file and the metadata file, so not sure what exactly is wrong. Honestly, I’m so lost and I don’t know what I’m doing wrong or where to go. Would really appreciate the help. Thank you in advance. ---
**Author:** ![jorondo1](https://yyz1.discourse-cdn.com/flex027/user_avatar/forum.biobakery.org/jorondo1/32/1312_2.png) [@jorondo1](https://forum.biobakery.org/u/jorondo1)\ **Post date:** [March 16, 2022, 1:49pm UTC](https://forum.biobakery.org/t/humann-3-0-function-downstream-analysis/3311/2 "2022-03-16T13:49:17Z")
I’m not sure if that’s what’s causing the negative entries (I don’t see how!) but I think for input into MaAsLin we need to normalize the tables differently using the _–units relab_ option. From the humann readme: _Prior to nomalization, select the scheme to use (copies per million or relative abundance). For example, if using MaAsLin, select relative abundance._ ---
**Author:** ![manzoor](https://avatars.discourse-cdn.com/v4/letter/m/77aa72/32.png) [@manzoor](https://forum.biobakery.org/u/manzoor)\ **Post date:** [January 13, 2023, 2:10am UTC](https://forum.biobakery.org/t/humann-3-0-function-downstream-analysis/3311/3 "2023-01-13T02:10:26Z")
> [@sarahi](#): > > Hi, I’m sorry for asking such a rudimentary question, but I’m really new to this. Basically, I’ve run my samples through HUMANN 3.0, and I’d like to do further analysis on the functional data; specifically the pathway abundances, and I’m completely lost on what to do next and what tools to use. Maybe I’m overthinking it but reading the descriptions of all the tools have confused me more. I have joined all my samples together in one table, and I tried to put it through MaAsLin 2 and I get this output: > > ```auto > Error in FUN(newX[, i], ...) : invalid 'type' (character) of argument > In addition: Warning message: > In vegan::decostand(features_norm, method = "total", MARGIN = 1, : > input data contains negative entries: result may be non-sense > > ``` > > I made sure that the sample columns lined up for both the function file and the metadata file, so not sure what exactly is wrong. > > Honestly, I’m so lost and I don’t know what I’m doing wrong or where to go. Would really appreciate the help. Thank you in advance. Hello, I am analysing metagenome data. We have around 300 samples including Patient and control groups. I have run HUMAnN3 and have individual sample outputs (gene family, path abundance and path coverage). Now, I want to merge them into one file (which support for down stram analyisis in R?). Basically, I want to see the differentially abundant pathways and gene families in the patient and control groups from the HUMAnN3 output. Which R package or tool I can use for this? I’m pretty new to this data analysis and was wondering if you can help. Thanks for your time, Muhammed ---
**Author:** ![Herve](https://yyz1.discourse-cdn.com/flex027/user_avatar/forum.biobakery.org/herve/32/3271_2.png) [@Herve](https://forum.biobakery.org/u/Herve)\ **Post date:** [January 16, 2023, 12:48pm UTC](https://forum.biobakery.org/t/humann-3-0-function-downstream-analysis/3311/4 "2023-01-16T12:48:44Z")
If you already normalized the tables with the --unit relab option, you can then merge them with the humann\_join\_tables function. The resulted tables can be imported into R. There, you can use Maaslin2 (or other differential abundance analysis tools) to figure out whether you observe differences between your patient and control groups, you simply need an additionnal dataset with your metadata. Significantly different features can be vizualised with heatmaps (produced by Maaslin or other R packages) or barplots (see ggplot2).