Details
-
Type:
Task
-
Status: Closed (View Workflow)
-
Priority:
Major
-
Resolution: Done
-
Affects Version/s: None
-
Fix Version/s: None
-
Labels:None
-
Story Points:2
-
Epic Link:
-
Sprint:Spring 6 2023 Mar 20, Spring 7 2023 Apr 10
Description
Muday Lab has detected possible sample switching in their time course data from three genotypes: ARE (mutant), F3H-OX3 (transgenic), and VF36 (wild-type, parent strain for F3H-OX3)
It was easy to identify which samples were ARE and not VF36 or F3H-OX3.
We need a way to distinguish F3H-OX3 from VF36 samples due to the sample switching issue.
To do this, we can use the fact that F3H-OX3 contains transgenic construct and VF36 does not.
Previously, we have found that the plant selective marker gene was expressed in soybean seeds in a transgenic line.
For this task, we'll investigate whether we can use the F3H-OX3 line's plant selection marker gene to distinguish transgenic from non-transgenic lines.
See:
- Open access article link: https://bmcbiotechnol.biomedcentral.com/articles/10.1186/s12896-015-0207-z
- Code repository: https://bitbucket.org/lorainelab/soyseq
Maarten has sent us three files with the construct information in them. These are added to the git repository here:
- ExternalDataSets/pK7WG2-F3H.fa
- ExternalDataSets/pK7WG2-F3H.gb
In this experiment, the plant selection gene was a kanamycin resistance gene.
For this task, use the kanamycin gene either as a query or a target in a search of the fastq files to identify samples that contain RNA-Seq reads from the kanamycin-resistance gene.
To do this, we need to investigate the different tools that might be available for this task.
Let's try to find an easy-to-use tool or approach that will let us identify RNA-Seq samples containing "kan" gene sequence.
Attachments
Issue Links
- relates to
-
-
- Closed
-
Activity
Some other notes related to this.
On Tuesday's tech meeting, the group discussed other possibilities that we could collectively try.
In the end, not much came up as things we can try.
Gloria is hunting down the original plasmid that was used in the transformation. Maybe from that there is some sequence we can also hunt for.
I talked about this with Rick White. He has done plant transformations and was fascinated by the problem. (And impressed how IGB was able to detect the mislabeling by the read count profiles!!). He suspected that we would not be able to detect kanamycin. Maybe in genome sequence but that too would be a long shot.
During discussion in scrum we realized that results are inconclusive from the kanamycin grep strategy. Let's halt this approach and try the approach from the soybean paper mentioned above. Making a new Jira issue to pursue this other strategy.
Closing this "investigate" ticket to try the other strategy.
The STAR run to see if ANY read sequences align to the kanamycin gene that Anthony from Gloria's lab sent.
He sent it in a Word doc: (KANR2)
GTCTGGAAAGAAATGCATAAACTTTTGCCATTCTCACCGGATTCAGTCGTCACTCATGGTGATTTCTCACTTGATAA
CCTTATTTTTGACGAGGGGAAATTAATAGGTTGTATTGATGTTGGACGAGTCGGAATCGCAGACCGATACCAGGA
TCTTGCCATCCTATGGAACTGCCTCGGTGAGTTTTCTCCTTCATTACAGAAACGGCTTTTTCAAAAATATGGTATTGA
TAATCCTGATATGAATAAATTGCAGTTTCATTTGATGCTCGATGAGTTTTAACATGGATGCTGATTTATATGGGTAT
AAATGGGCTCGCGATAATGTCGGGCAATCAGGTGCGACAATCTATCGCTTGTATGGGAAGCCCGATGCGCCAGAG
TTGTTTCTGAAACATGGCAAAGGTAGCGTTGCCAATGATGTTACAGATGAGATGGTCAGACTAAACTGGCTGACG
GAATTTATGCCTCTTCCGACCATCAAGCATTTTATCCGTACTCCTGATGATGCATGGTTACTCACCACTGCGATCCCC
GGAAAAACAGCATTCCAGGTATTAGAAGAATATCCTGATTCAGGTGAAAATATTGTTGATGCGCTGGCAGTGTTCC
TGCGCCGGTTGCATTCGATTCCTGTTTGTAATTGTCCTTTTAACAGCGATCGCGTATTTCGTCTCGCTCAGGCGCAA
TCACGAATGAATAACGGTTTGGTTGATGCGAGTGATTTTGATGACGAGCGTAATGGCTGGCCTGTTGAACAAGTC
TGGAAAGAAATGCATAAACTTTTGCCATTCTCACCGGATTCAGTCGTCACTCATGGTGATTTCTCACTTGATAACCT
TATTTTTGACGAGGGGAAATTAATAGGTTGTATTGATGTTGGACGAGTCGGAATCGCAGACCGATACCAGGATCT
TGCCATCCTATGGAACTGCCTCGGTGAGTTTTCTCCTTCATTACAGAAACGGCTTTTTCAAAAATATGGTATTGATA
ATCCTGATATGAATAAATTGCAGTTTCATTTGATGCTCGATGAGTTTT
I made a fasta out of this and is on the cluster at this location:
/nobackup/tomato_genome/muday-144/nfcore-2022/kanhunt/kanR2_SeqfromAnthony-fullsequence.fna
I treated the above sequence as the reference file for a STAR alignment.
Results to go here:
/nobackup/tomato_genome/muday-144/nfcore-2022/kanhunt/star
First step in STAR is the make a genome index file.
This is done once via this script:
/nobackup/tomato_genome/muday-144/nfcore-2022/kanhunt/star-generategenome.slurm
And then we run 144 jobs as an array, 1 for each read sequence in the experiment:
star-chunks.slurm
(silly name, I first planned to align to small chunks, named it this way)
We get 144 bam files, all of which are empty. The log files appear that STAR ran successfully.
Blast RUN:
Just like Star run but for blast.
Location:
/nobackup/tomato_genome/muday-144/nfcore-2022/kanhunt/blast
Script:
/nobackup/tomato_genome/muday-144/nfcore-2022/kanhunt/
-Need to run seqtk to convert the fastq to fasta to run blast.
So far there is an issue. 35 jobs complete. And the rest do not.
Getting kicked off the cluster because of my internet. Hope to resolve from office.
Of the 35 jobs that completed, no sequences were found to hit at a e-score cutofff = 1e-25.
Using kan-r gene in this way does not yield much. Closing this ticket.
Next strategy to try will be running an RNA-Seq alignment tool on the fastq files and the new plasmid genome, either by itself or with the rest of the tomato SL5 genome assembly, using the fasta file and bed file made by [~molly] in IGBF-3306.
We will make a new ticket for running this RNA-Seq alignment strategy to detect Kan-r expression.
Ann's first idea on what to do:
Augment the tomato SL5 genome assembly and gene model annotations by adding a new "fake" chromosome containing the construct sequence and construct annotations provided by Maarten. Then, we use this as the target assembly in a fresh run of the nf-core rnaseq pipeline. This should produce a new "counts" file with counts for the newly added construct gene locations, including counts for the kanamycin resistance gene. This might not be a great idea! We might be able to find a more direct and simple approach that would be less work. But if we can't, I'm pretty sure this would get the job done.