Subio Platform allows you to quantify gene and transcript expression directly from RNA-Seq FASTQ files and analyze the results. The entire process can be performed on a standard Windows PC or Mac, without having to use the UNIX command line.
You can choose between two methods for expression quantification: the fast Kallisto method and HISAT2 + StringTie, which maps reads to a reference genome. If processing is interrupted or Subio Platform terminates unexpectedly, you can resume processing without repeating completed steps by reusing previously saved results.
This article explains how to choose between the two methods, process FASTQ files, and import the resulting expression data into Subio Platform.
Kallisto or HISAT2 + StringTie: Which Should You Use?
Subio Platform supports two methods for quantifying RNA-Seq expression: Kallisto and HISAT2 + StringTie.
| Kallisto | HISAT2 + StringTie | |
|---|---|---|
| Method | Estimates expression using known transcript sequences | Maps reads to a reference genome to estimate expression |
| Processing Speed | Very fast | Slower than Kallisto |
| Reference Data | Transcriptome | Genome and gene annotations |
| Main Applications | Expression analysis of known genes and transcripts | The same applications as Kallisto, plus analyses requiring BAM files or genome alignment information |
Kallisto Is Recommended for Routine Gene Expression Analysis
Kallisto rapidly estimates transcript expression levels by matching RNA-Seq reads against a reference transcriptome. Because it does not perform conventional alignment that determines the genomic position of each read, it can greatly reduce processing time and memory requirements. For expression comparisons, clustering, and other analyses based on the expression levels of known genes or transcripts, we recommend starting with Kallisto.
Use HISAT2 + StringTie When Genome Alignment Information Is Needed
HISAT2 maps RNA-Seq reads to a reference genome. StringTie then uses the alignment results to estimate expression levels. Although this method requires more processing time and storage space than Kallisto, it is suitable when BAM files or genome alignment information are needed for further analysis.
With either method, the calculated expression values can be imported into Subio Platform for analysis.
Before You Start: Preparing Tools and Reference Data
Before processing FASTQ files, you need to prepare the required tools and reference data. If this is your first time using the feature, follow the setup instructions for your operating system.
Once the setup is complete, follow the instructional videos below to process your FASTQ files.
Processing FASTQ Files with Kallisto
[Kallisto instructional video to be inserted here]
To use Kallisto, prepare the Kallisto executable
and a reference transcriptome for the species you are analyzing.
Select Kallisto in Subio Platform,
configure the tool and reference data,
and specify the FASTQ files to process.
gzip-compressed FASTQ files can be used directly
in .fastq.gz format without decompression.
Processing proceeds automatically, from quality control and filtering with fastp to expression quantification with Kallisto. Once processing is complete, estimated Counts and TPM values are available at both the gene and transcript levels. Finally, import the calculated expression values into Subio Platform.
When importing the data, you can choose either Gene ID or Gene Symbol. Because multiple Gene IDs may correspond to the same Gene Symbol, we recommend selecting Sum to combine values associated with the same Gene Symbol when importing by Gene Symbol. Count values are stored as Ch1 Raw Signal, while TPM values are stored as Ch1 Reserved.
Processing FASTQ Files with HISAT2 + StringTie
[HISAT2 + StringTie instructional video to be inserted here]
To use HISAT2 + StringTie, prepare both tools, a reference genome, and gene annotations for the species you are analyzing. Select HISAT2 + StringTie in Subio Platform, configure the tools and reference data, and specify the FASTQ files to process. As with Kallisto, gzip-compressed FASTQ files can be used directly.
Processing proceeds through quality control and filtering with fastp, read alignment with HISAT2, and expression quantification with StringTie. The resulting expression values can be imported into Subio Platform in the same way as those obtained with Kallisto.
Pausing, Resuming, and Handling Processing Errors
Processing RNA-Seq FASTQ files can take considerable time, depending on the amount of data and your computer's performance. Subio Platform allows you to pause or stop processing using Pause or Stop. If an error occurs during processing, you can also resume by reusing the results of completed steps, without having to restart the entire process from the beginning. When resuming, Subio Platform checks which steps have been completed for each sample and runs only the necessary steps.
Do Not Delete Folders Created During Processing
When processing is interrupted or resumed, multiple folders containing intermediate results may be created. These folders contain the results of steps that have already been completed. If you may need to resume processing, do not delete any folders or files created during processing.
Subio Platform reuses valid saved results and continues processing from the necessary step. When analysis completes successfully, the final results are organized into a single folder. There is no need to worry if additional folders appear during processing.
When an Error Occurs
If an error occurs during processing, Subio Platform automatically retries the failed operation. If necessary, it may reduce the number of threads and continue processing. If a FASTQ file itself is invalid, the file can be excluded so that processing can continue for the remaining samples.
However, not all errors can be resolved automatically. In some cases, you may need to identify the cause, correct the settings or FASTQ files, and run the process again.
Where Are the Analysis Results Saved?
Analysis results are saved in a folder named
rna-seq-pipeline.date-time-UUID.
The paths below are relative to this folder.
sample_name and UUID
are placeholders for the actual sample names and identifiers.
The .gz extension indicates a gzip-compressed file.
| File or Folder | Description |
|---|---|
merged.tsv.gz
|
A combined expression table containing gene-level Counts and TPM values for all samples. |
pre-alignment/sample_name.html
pre-alignment/sample_name.json
|
Quality control and filtering results from fastp. The HTML file is a report that can be viewed in a browser, while the JSON file contains summary statistics. |
alignment/sample_name.bam
alignment/sample_name.bai
|
HISAT2 pipeline only. Alignment results and their index. The BAM file can also be used for reanalysis with StringTie. |
alignment/sample_name.txt
|
HISAT2 pipeline only. A summary of the alignment results. |
alignment/UUID.abundance.tsv.gz
|
Kallisto pipeline only. Original transcript-level quantification results from Kallisto, including transcript length, effective length, estimated Counts, and TPM. These files can be used for reanalysis with tximport. |
alignment/UUID.run_info.json
|
Kallisto pipeline only. Run information, including the number of processed reads and mapping statistics. |
post-alignment/sample_name_gene.txt.gz
|
Gene-level Counts and TPM values for each sample, including Gene ID and Gene Symbol. |
post-alignment/sample_name.excluded-transcripts.tsv
|
Kallisto pipeline only. Transcript IDs that could not be uniquely matched to genes in the GTF and were excluded from gene-level aggregation, along with their estimated Counts, TPM values, and totals. |
summary.txt
|
A summary of filtering, alignment, and quantification results for each sample. |
versions.log
|
Versions of the tools used and analysis settings. |
run.log.gz
|
Execution log containing processing progress, external commands, errors, and other information. During processing or before compression, the file is named run.log. |
post-alignment/sample_name_transcript.txt.gz
|
Transcript-level Counts and TPM values for each sample, including Transcript ID, Gene ID, and Gene Symbol. |
Merged_transcript_level.txt.gz
|
A combined expression table containing transcript-level Counts and TPM values for all samples. |
Additional JSON files and other files used for resuming processing
and managing results are also saved.
If previous runs exist from before a retry or resumption,
their records are retained in the history folder.
StringTie output GTF files are removed during cleanup after successful completion.
Some expression tables and log files are gzip-compressed to reduce storage requirements. The analysis results can be used not only for importing into Subio Platform, but also for further analysis with external tools such as R.
Check the Execution Log If Processing Fails
If an error prevents processing from completing successfully,
check run.log.
This log records information about the tools and analysis settings used,
processing progress, errors, and other details.
If the log has been compressed as run.log.gz,
decompress it before viewing its contents.
If you cannot identify the cause of an error, we recommend uploading run.log to an AI assistant such as ChatGPT and asking it to investigate the cause and suggest possible solutions. For example, you can use the following prompt:
I am processing RNA-Seq FASTQ files using Subio Platform, but the process stopped because of an error. Please examine the attached run.log file and identify the processing step where the error occurred, possible causes, and what I should check to resolve the problem.
AI can help you find relevant error messages in the log and make troubleshooting more efficient. However, the log alone may not be sufficient to identify the cause, and AI-generated answers are not always correct. Review any suggested solutions before applying them. Before uploading log files to an external AI service, make sure they do not contain personal or confidential information.
Recalculating and Analyzing Expression Data with External Tools
Expression values calculated by Subio Platform can be analyzed not only within Subio Platform, but also with external tools such as R. If you want to use a different method for gene-level aggregation, you can also recalculate the expression values using files saved in the final results folder.
For Kallisto
Subio Platform calculates gene-level expression values
by separately summing the estimated Counts and TPM values
of transcripts belonging to the same gene.
To reaggregate the results using tximport,
use each sample's *.abundance.tsv.gz file
saved in the alignment folder,
together with a table mapping Transcript IDs to Gene IDs
based on the GTF used for the analysis.
For HISAT2 + StringTie
To reaggregate the results using tximport,
rerun StringTie with the -e -B options,
using the BAM files saved in the alignment folder
and the reference GTF used for the analysis.
The resulting Ballgown-format quantification files can be imported into tximport.
There is no need to repeat the alignment step from the FASTQ files.
These methods allow you to apply different aggregation methods or analysis settings from those used for the expression values imported into Subio Platform.
For the subsequent analysis steps, see Chapter 2 and later of RNA-Seq Data Analysis Tutorial: Learning the Analysis Workflow and How to Interpret Results Using GEO Data .