Frequently Asked Questions¶
Qiita data disclaimer¶
Qiita is a research tool, and as such, is hosted on research computing resources maintained by the Knight Lab at the University of California San Diego.
Data privacy is a key aspect of our operations, and is strictly adhered to at every step of the workflow. We are committed to protecting any and all information (including sequence data) submitted to Qiita. For example, your data is sandboxed by default upon upload, and remains private at the discretion of the Owner (i.e., you are the Owner) of the study.
Authorizations and access associated with any given study is maintained and controlled by the Owner of the study; importantly, this means that sharing rights of a study within Qiita is determined solely by the Owner.
A study within Qiita, and its associated sequence and metadata, can be permanently deleted by the Owner as long as it is not public.
What kind of data can I upload to Qiita for processing?¶
Processing in Qiita requires 3 things: raw data, sample and prep information files. Here you can find a list of currently supported raw files files. Note that we are accepting any kind of target gene (16S, 18S, ITS, whatever), Whole Genome and Metatranscriptomic. Check our Processing recommendations.
What’s the difference between a sample and a prep information file?¶
A sample information file describes the samples in a study, including environmental factors relating to the associated host. The prep information file has information on how the sample was processed in the wet lab. If you collected 100 samples for your study, you will need 100 rows in your sample information file describing each of them, and additional rows for blanks and other control samples. If you prepared 95 of them for 16S and 50 of them for 18S, you will need 2 prep information files: one with 95 rows describing the preparation for 16S, and another one with 50 describing the 18S. For more information visit Qiita allows for complex study designs.
How should I split my samples within preparations?¶
This question normally comes up when you are working with per sample FASTQs as at this stage there is no lane and run grouping within the samples.
Generally, we recommend to set a single preparation for each lane in each sequencing run. This separation allows users to first test there are no biases within their lanes and runs, and optionally merge them in a single analysis. For your convenience, when you create a new analysis you can keep samples separate (default) or merge samples with matching names in different preparations.
Another thing to consider is that once you are ready, you might want to submit to EBI-ENA and they have a limit of 10M on the metadata that you want to submit. This limitation is a combination of the metadata of each preparation and the sample information. For example, if we imagine the worst case scenario, that we have in all our columns the default ‘not applicable’ NULL value, and that XML adds 4 times the size due to its formatting (how EBI-ENA expects the submission); then we will have len(‘not applicable’) * 50 * 4 = 2800 (2.8K) per sample. Thus, we cannot have more than 3.5K samples. Note that this number depends on the number of metadata columns in your sample information file and the number of characters in the values of each sample and metadata column.
Please do not hesitate to send us an email if you have questions about this.
Example study processing workflow¶
A few more instructions: for the example above the workflow should be:
Create a new study.
Add a sample information file. You can add 1, try to process it and the system will let you know if you have errors or missing columns. The most common errors are: the sample name column should be named sample_name, duplicated sample names are not permitted. For a full list of required fields, visit Getting Started Guide.
Add a prep information file to your study for each data type. The prep information file should contain all the samples in the sample information file or a subset. If you have more than one FASTQ file set (forward, reverse (optional) and barcodes) you will need to add a run_prefix column, see Prepare information files. A prep information file and a QIIME compatible mapping file will be available for download after the prep information file is added successfully.
Upload and link your raw data to each of your prep information files. Depending on your barcoding/sequencing strategy you might need 1 or more raw data file sets. If you have 2 raw data sets you may have to rename one set so that each set has a different name. If they have the same name they will over-write on upload. Note that you can have one FASTQ file set linked to more than one prep information file.
Preprocess your files. For target gene amplicon sequencing, this will demux and QC. There are multiple options for preprocessing depending on the barcode format and the data output from the sequencing center - this may require a series of trial and error to establish the correct option for your data files. After demultiplexing a log file is generated with statistics about the files demultiplexed including the number of sequences assigned per sample.
Process each of your preprocessed data types. For target gene, this will perform closed OTU picking against the latest version of Greengenes and can be quite time consuming depending on the number of samples and the depth of sequencing.
How to solve download or unzip errors?¶
Dealing with large files might be daunting but, in general, following these instructions should make things easier. First, make sure that you have enough space for the zip download file; if you are unsure of the size required click on the button and your browser will show an estimate size of the download. Second, make sure that your computer has all the sleep settings turned off; for example, in a Mac, got to System Preferences, Energy Saver, Power Adapter and unselect the option of “Put hard disks to sleep when possible”; don’t forget to save the settings. Third, download the file but point to the storage that you want to save your file in; using Chrome, right click on the download button and select “Save Link As …”; and select the location where you have enough space (see point 1). Fourth, wait for the download to finish, this will depend on your Internet service. Finally, unzip the file with a newer version of zip (see below).
By the way, if you are a developer and would like to add to Qiita the possibility of resumable downloads, we would happily welcome this contribution.
Now, when trying to open the large downloaded zip file there is a change that you will get an error like: “start of central directory not found; zipfile corrupt”. This issue arises from using old versions of zip and you need to have unzip >= 6.0.0. To check you unzip version you can run: unzip -v.
To update your unzip for most operating systems you can simply use your regular package admin program. However, for Mac we suggest using this version of unzip.
Additionally, there is a chance that you will see an error/warning message like this:
extracting: BIOM/57457/all.biom bad CRC f6b2a86b (should be 38903659). These
messages are consequence of the zip library we are using internally and are fine to
ignore. If you want to check them, we suggest taking any of the files and generating their
CRC32 checksum; in MAC’s you can run
crc32 [filename] and should get the first number
in that message; for example:
$ crc32 57457_all.biom f6b2a86b
Do you have specific Qiita questions?¶
Please send them to firstname.lastname@example.org; this will assure prompt responses while keeping your information private.
Do you have general analytical questions?¶
Normally these are: How can I test X factor in my samples? Why do I see this pattern? Which statistical method better fits my question?
As you can imagine, you are not alone as this is a common problem while doing analysis. Thus, we suggest posting your data processing questions (raw-data to feature-table) to email@example.com and general-interest analytical questions (feature-table analyses) to the QIIME2 Forum, please be sure to add your question in the “General Discussion” category of the forum. This will generally ensure that your question is answered in a timely manner. There are many users and developers monitoring the QIIME2 Forum. Posting questions in the forum allows you to share answers with others, who may have similar questions in the future.
Do you have Meta-Analysis questions?¶
A common thing is why do I have a given pattern in my analysis, like PCoA plots or taxonomy summaries.
Let’s start by saying, this is an open area of research so we are still learning about the effect sizes and what matters in individual and meta-analysis. However, there are a few good resources to help you understand those patterns:
How to solve BIOM name errors?¶
When uploading a BIOM table, you may get an error like: “The sample ids in the BIOM table do not match the ones in the prep information. Please, provide the column “run_prefix” in the prep information to map the existing sample ids to the prep information sample ids.”. This issue arises if your sample names in your BIOM table do not match with the sample names in your preparation information file.
To correct this issue, simply add a column to your preparation information file named “run_prefix”. In this column, add the sample names from your BIOM table that matches the sample names listed in the sample_name column in your preparation information file.
As a reminder, direct BIOM uploads cannot become public in the system.
What’s a Qiita Artifact?¶
A Qiita artifact is a collection of files and their summaries that represent the output or input of a processing or analytical command.
For example a per_sample_FASTQ artifact will contain the per sample FASTQ files and their summary (if a user generated); while a BIOM artifact has the feature table as a biom file, a QIIME2 QZA, any other supporting files (like a phylogenetic tree for deblur or sortmerna_picked_otus.tgz for close reference picking), and summaries.
How to convert Qiita files to QIIME2 artifacts?¶
How to add extra files to a Qiita study?¶
Many publications rely on extra files that are not part or generated within Qiita. However, to facilitate analytical reproducibility a user might like to link these files to Qiita. In this case, we recommend to upload your external file to a long term repository, like figshare.com, and then link to your study via the “Analytical Notes” within a study. The “Analytical Notes” section can be accessed within the study “Edit” button. Note that this text box renders Markdown when displayed in the study section. Markdown allows to format text, add images, etc; for more information check this 3 minute read about Markdown.
Where’s my QIIME1 mapping file?¶
During the 2020.11 deployment we removed the functionality that automatically created the merged preparation and sample information file per preparation. This change will allow us to make faster information file updates allow for future multi-site operations.
If you want to create a merged and validated mapping file (merged sample and preparation information file) please create an analysis by following these instructions: Creating A New Analysis.
I want to transfer a lot of files to Qiita, is there an easy way?¶
Yes! This is available in the “Upload Files” section of each study by accessing the tab “Upload via Remote Server (ADVANCED)”.
We currently suggest using scp, note that your server where you store your files should allow connecting to it via scp - in other words, this only works when moving files from a server to Qiita.
Now, the way it works is that you need to create a new secure key to that server, imagine that you are making a copy of your storage-shed key, then you share that key to Qiita (you will give access to Qiita to run a single copy command in your server), Qiita uses that key and securely destroy it.
To take advantage of this feature you need to:
Prepare all the files you want to transfer in a single folder (or you can transfer multiple times from multiple folders into one study). Note that Qiita will only collect files with valid extensions from the top level directory (no sub-directories); see the top of the “Upload Files” page within your study for the latests list of valid extensions.
In your server (this needs to be run within your home directory in the server where you store the files!), generate a new key by running: ssh-keygen -t rsa -C “ssh test key” -f ~/.ssh/qiita-key -P “”. Here is where you are creating that key to your storage-shed.
Allow access using the new key to new connections (this also needs to be run in the remote server): cat ~/.ssh/qiita-key.pub >> ~/.ssh/authorized_keys. This tells the server that is OK to give access to the key created to your storage-shed; note that if you want to completely stop that key to work you can open that file and remove the line with the name of this key.
Dowload your new generated key qiita-key (the file) to your local computer and use it in the Key option of “Upload via Remote Server (ADVANCED)”.
Using this key you can List Files to test the connection and verify the list of study files. Then, if the connection is made and files are correct, press ‘Transfer Files’ to initiate the transfer.
Note that if you click multiple times, too quickly there is a chance that your server will block Qiita, if this happens, just wait a few minutes and retry again.
How do I update the sample or preparation file?¶
Remember, these are separate files so they need to be updated separately. In both cases, the easiest is to upload the new file from your computer to Qiita using the Upload Files button in your study. Once the file you want to use is there you can use them within Qiita.
To update a sample information file: Click on Sample Information button in the your study page, then use the Update sample information section on that page to select your file and update it. Note that for this you can also directly upload your file via the Direct upload file (< 2MB).
To update a preparation information file: Click on the preparation you want to update within your study page, then click on Summary, and use the Update prep information section on that page to select your file and update it.
Note that these information is generally independent of the sequence processing so you don’t need to reprocess your sequences; however, if you use your study in an analysis, you will need to recreate that analysis to use the updated sample or preparation metadata
When do I need the run_prefix in my preparation information file?¶
It depends on your sequence processing but in general it will facilitate loading your files to your preparation in Qiita.
First of all, this is a prefix value so you only need the beginning of the file name to load the files in Qiita; for example if your file names for a given sample are: AWERWADFA_I1.fastq.gz, AWERWADFA_R1.fastq.gz, AWERWADFA_R2.fastq.gz, the run_prefix for the sample should be AWERWADFA. Qiita will use that to group those files under the same sample with that run_prefix.
Now the run_prefix is used constantly within file selection and processing in all file types but specially on:
BIOM: the run_prefix is used to rename the samples in your BIOM to match the sample names in Qiita. Basically, if you add the sample names in your BIOM file as the run_prefix and the sample name in Qiita in the sample_name column of your preparation, Qiita will automatically rename them to match.
per sample FASTQ: run_prefix is the way to link which sample goes with which files so using here will facilitate loading your files to the preparation and then used for processing, without it Qiita will not be able to process your samples.
How to cite Qiita?¶
If you use Qiita for processing, submission to EBI-ENA and/or its data for any published research, please include the following citation:
Qiita: rapid, web-enabled microbiome meta-analysis. Antonio Gonzalez, Jose A. Navas-Molina, Tomasz Kosciolek, Daniel McDonald, Yoshiki Vázquez-Baeza, Gail Ackermann, Jeff DeReus, Stefan Janssen, Austin D. Swafford, Stephanie B. Orchanian, Jon G. Sanders, Joshua Shorenstein, Hannes Holste, Semar Petrus, Adam Robbins-Pianka, Colin J. Brislawn, Mingxun Wang, Jai Ram Rideout, Evan Bolyen, Matthew Dillon, J. Gregory Caporaso, Pieter C. Dorrestein & Rob Knight. Nature Methods, volume 15, pages 796–798 (2018); https://doi.org/10.1038/s41592-018-0141-9.