Last updated: 2024-02-19

Checks: 7 0

Knit directory: locust-comparative-genomics/

This reproducible R Markdown analysis was created with workflowr (version 1.7.1). The Checks tab describes the reproducibility checks that were applied when the results were created. The Past versions tab lists the development history.

R Markdown file: up-to-date

Great! Since the R Markdown file has been committed to the Git repository, you know the exact version of the code that produced these results.

Environment: empty

Great job! The global environment was empty. Objects defined in the global environment can affect the analysis in your R Markdown file in unknown ways. For reproduciblity it’s best to always run the code in an empty environment.

Seed: set.seed(20221025)

The command set.seed(20221025) was run prior to running the code in the R Markdown file. Setting a seed ensures that any results that rely on randomness, e.g. subsampling or permutations, are reproducible.

Session information: recorded

Great job! Recording the operating system, R version, and package versions is critical for reproducibility.

Cache: none

Nice! There were no cached chunks for this analysis, so you can be confident that you successfully produced the results during this run.

File paths: relative

Great job! Using relative paths to the files within your workflowr project makes it easier to run your code on other machines.

Repository version: 55f9385

Great! You are using Git for version control. Tracking code development and connecting the code version to the results is critical for reproducibility.

The results in this page were generated with repository version 55f9385. See the Past versions tab to see a history of the changes made to the R Markdown and HTML files.

Note that you need to be careful to ensure that all relevant files for the analysis have been committed to Git prior to generating the results (you can use wflow_publish or wflow_git_commit). workflowr only checks the R Markdown file, but you know if there are other scripts or data files that it depends on. Below is the status of the Git repository when the results were generated:


Ignored files:
    Ignored:    .DS_Store
    Ignored:    .RData
    Ignored:    .Rhistory
    Ignored:    .Rproj.user/
    Ignored:    analysis/.DS_Store
    Ignored:    analysis/.Rhistory
    Ignored:    data/.DS_Store
    Ignored:    data/.Rhistory
    Ignored:    data/metadata/.DS_Store
    Ignored:    figures/

Note that any generated files, e.g. HTML, png, CSS, etc., are not included in this status report because it is ok for generated content to have uncommitted changes.

These are the previous versions of the repository in which changes were made to the R Markdown (analysis/3_seq-data-qc.Rmd) and HTML (docs/3_seq-data-qc.html) files. If you’ve configured a remote Git repository (see ?wflow_git_remote), click on the hyperlinks in the table below to view the files as they were in that past version.

File	Version	Author	Date	Message
Rmd	55f9385	Maeva A. TECHER	2024-02-19	wflow_publish("analysis/3_seq-data-qc.Rmd")
html	df94db2	Maeva A. TECHER	2024-02-19	adding markdown qc
Rmd	f6b4961	Maeva A. TECHER	2024-02-19	wflow_publish("analysis/3_seq-data-qc.Rmd")
html	e39d280	Maeva A. TECHER	2024-01-29	Build site.
html	f701a01	Maeva A. TECHER	2024-01-29	reupdate
html	be046c6	Maeva A. TECHER	2024-01-24	Build site.
html	1b09cbe	Maeva A. TECHER	2024-01-24	remove
html	141b63c	Maeva A. TECHER	2023-12-18	Build site.
Rmd	53877fa	Maeva A. TECHER	2023-12-18	add pages

At this point, we will have organized both SRA and de novo paired-end sequencing data within our reads directory located at /scratch/group/songlab/maeva/headthor-locusts-rna/{species}/data/reads and will be ready to run the Snakemake pipeline on it. As a reminder, each library corresponds to the transcriptome of bulk tissues (either head or thorax) from a single specimen reared under isolated or crowded conditions.

Load R libraries (install first from CRAN or Bioconductor)

library("knitr")
library("rmdformats")
library("tidyverse")
library("DT")  # for making interactive search table
library("plotly") # for interactive plots
library("ggthemes") # for theme_calc
library("reshape2")
library("readr")
library("ggplot2")

## Global options
options(max.print="10000")
knitr::opts_chunk$set(
    echo = TRUE,
    message = FALSE,
    warning = FALSE,
    cache = FALSE,
    comment = FALSE,
    prompt = FALSE,
    tidy = TRUE
)
opts_knit$set(width=75)

Control the quality of the .fastq files

The first step is to ensure that we have enough reads per library and remove any potential outlier resulting from library preparation or sequencing failure. For that, we will assess each .fastq file with FASTQC. However, considering we are working with a significant sample size, we will compile the results using MULTIQC.

The aggregated results can be conveniently viewed by opening the HTML report in a web browser.

module load GCC/12.2.0  OpenMPI/4.1.4 MultiQC/1.14
multiqc --title 'TYPE THE TITLE YOU WANT' -v /PATHTODIRECTORY

montana <- read_table("data/metadata/Stats_RNAseq_QC_19Feb2024.txt", col_names  = TRUE,  guess_max = 1000) 
head(montana)

FALSE # A tibble: 6 × 7
FALSE   Sample_Name      Species Perc_Dups Perc_GC M_Seqs Unique_Reads Duplicate_Reads
FALSE   <chr>            <chr>       <dbl>   <dbl>  <dbl>        <dbl>           <dbl>
FALSE 1 SAMER_G_Crd_SRR… americ…      0.71    0.46   23.7      6958831        16765516
FALSE 2 SAMER_G_Crd_SRR… americ…      0.67    0.47   23.7      7784916        15939431
FALSE 3 SAMER_G_Crd_SRR… americ…      0.71    0.48   15.7      4570667        11085054
FALSE 4 SAMER_G_Crd_SRR… americ…      0.67    0.48   15.7      5145265        10510456
FALSE 5 SAMER_G_Crd_SRR… americ…      0.84    0.49   37.2      6049864        31125514
FALSE 6 SAMER_G_Crd_SRR… americ…      0.8     0.49   37.2      7301268        29874110

# Convert values to millions
montana <- montana %>%
  mutate_at(vars(contains("Reads")), list(~ ./1000000))

# Pivot the data
montana_long <- montana %>%
  pivot_longer(cols = contains("Reads"), names_to = "Variable", values_to = "Value")

# Define colors
colors.reads <- c("Duplicate_Reads" = "black", "Unique_Reads" = "deepskyblue")

# Plot the stacked bar plot with values in millions and custom colors
ggplot(montana_long, aes(x = Sample_Name, y = Value, fill = Variable)) +
  geom_bar(stat = "identity", position = "stack") +
  facet_wrap(~Species, scales = "free") +
  labs(title = "Stacked Bar Plot of Unique Reads and Duplicate Reads by Sample",
       x = NULL,   # Remove x-axis label
       y = "Millions of Reads") +
  theme_minimal() +
  theme(axis.text.x = element_blank()) +  # Remove x-axis labels
  scale_fill_manual(values = colors.reads) +  # Set custom colors
  geom_hline(yintercept = 30, linetype = "dashed", color = "red") +
  geom_hline(yintercept = 50, linetype = "dashed", color = "gray70") +
  ylim(0, 62)  # Set y-axis limits

library(ggConvexHull)
# Define custom colors for each species
species_colors <- c("americana" = "forestgreen", "cubense" = "yellow3", "gregaria" = "orange", "nitens" = "blue", "piceifrons" = "red2", cancellata = "deeppink")  

p <- ggplot(montana, aes(x = Perc_Dups, y = M_Seqs, color = Species)) +
  geom_point(size =2) +
  scale_color_manual(values = species_colors) +  # Apply custom colors
  labs(title = "M_Seqs vs % Dups by Species",
       x = "Percentage of Duplicates",
       y = "Millions of Reads") +
  theme_minimal()

p + geom_convexhull(aes(fill = Species, color = Species), alpha = 0.2) +
    scale_fill_manual(values = species_colors)

Trim and adapter removal

After checking the initial sequence quality, we can determine whether any parameters adjustments are needed for Trimmomatic. This is particularly pertinent for the ‘trailings’ and ‘leading’ parameters which removes the nucleotides at the end and start

########################################
# Snakefile rule
########################################

rule trim_adapt:
    input:
        read1 = OUTdir + "/reads/{locust}_1.fastq.gz",
        read2 = OUTdir + "/reads/{locust}_2.fastq.gz",
        adaptfile = OUTdir + "/list/TruSeqNextera_PE.fa"
    output:
        trimmedread1 = OUTdir + "/trimming/{locust}_trim1P_1.fastq.gz",
        badread1 = OUTdir + "/trimming/{locust}_trim1U_1.fastq.gz",
        trimmedread2 = OUTdir + "/trimming/{locust}_trim2P_2.fastq.gz",
        badread2 = OUTdir + "/trimming/{locust}_trim2U_2.fastq.gz",
    shell:
        """
        module load Trimmomatic/0.39-Java-11
        java -jar $EBROOTTRIMMOMATIC/trimmomatic-0.39.jar PE -threads 2 -phred33 {input.read1} {input.read2} {output.trimmedread1} {output.badread1} {output.trimmedread2} {output.badread2} ILLUMINACLIP:{input.adaptfile}:2:30:10 LEADING:30 TRAILING:30 SLIDINGWINDOW:4:15 MINLEN:36 
        """
        
########################################
# Parameters in the cluster.json file
########################################

    "trim_adapt":
    {
        "cpus-per-task" : 2,
        "partition" : "medium",
    "ntasks": 2,
        "mem" : "1G",
        "time": "0-04:00:00"
    },

SLIDINGWINDOW We use the 4:15 approach which means that in a 4-base window, if the average quality drops below 15, the bases will be trimmed.
MINLEN:36 Specifies the minimum length a read must be to be kept after all trimming steps is at 36 bp otherwise it is too short to convey any information.

We used the Illumina adapters for library preparation and as per recommended on the support website, we removed TruSeq.

We created a list called TruSeqNextera_PE.fa that gathered several adapters. >PrefixPE/1 TACACTCTTTCCCTACACGACGCTCTTCCGATCT >PrefixPE/2 GTGACTGGAGTTCAGACGTGTGCTCTTCCGATCT >PE1 TACACTCTTTCCCTACACGACGCTCTTCCGATCT >PE1_rc AGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGTA >PE2 GTGACTGGAGTTCAGACGTGTGCTCTTCCGATCT >PE2_rc AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC>PrefixNX/1 AGATGTGTATAAGAGACAG >PrefixNX/2 AGATGTGTATAAGAGACAG >Trans1 TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG >Trans1_rc CTGTCTCTTATACACATCTGACGCTGCCGACGA >Trans2 GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG >Trans2_rc CTGTCTCTTATACACATCTCCGAGCCCACGAGAC

Trim quality control

We always perform quality control checks after trimming to ensure that the clipping and filtering of sequences were not overly aggressive. Given the number of sequences we work with, manually inspecting each file immediately would be time-consuming. Instead, we adopt a random sampling approach across different species, rearing conditions, and tissues to see that the trimming process worked well.

########################################
# Snakefile rule
########################################

#Quality control step after trimming: checked for adapter content in particular and quality scores
rule trim_fastqc:
    input:
        read1 = OUTdir + "/trimming/{locust}_trim1P_1.fastq.gz",
        read2 = OUTdir + "/trimming/{locust}_trim2P_2.fastq.gz",
    output:
        htmlqc1 = OUTdir + "/trimming/{locust}_trim1P_1_fastqc.html",
        htmlqc2 = OUTdir + "/trimming/{locust}_trim2P_2_fastqc.html",
    shell:
        """
        module load FastQC/0.11.9-Java-11
        fastqc {input.read1}
        fastqc {input.read2}
        """
        
########################################
# Parameters in the cluster.json file
########################################

    "trim_fastqc":
    {
        "cpus-per-task" : 2,
        "partition" : "medium",
        "ntasks": 1,
        "mem" : "500M",
        "time": "0-03:00:00"
    },

Non-target sequencing data

Contamination is likely to occur throughout various stages of experiments, including tissue acquisition, RNA extraction and library preparation. One can hope to reduces as much as possible its impact on the final sequencing data which can affect the success rate of reads mapping.

Screen for microbial contamination

We decided to screen for microbes sequences present in the trimmed paired-end reads FASTQ.gz files using the tool Kaiju. Kaiju translates metatranscriptomics sequencing reads into six possible reading frames and searches for maximum exact matches of amino acid sequences in a given annotation protein database.

We used the most extensive microbial database nr_euk which encompass the subset of NCBI BLAST nr database containing all proteins belonging to Archaea, Bacteria, Viruses, Fungi and microbial Eukaryotes.

## Downloading the 2022-03-10 database from Kaiju webserver
wget https://kaiju.binf.ku.dk/database/kaiju_db_nr_euk_2022-03-10.tgz

To run Kaiju we used the following Snakemake rule:

########################################
# Snakefile rule
########################################

rule kaiju:
        input:
                read1 = OUTdir + "/trimming/{locust}_trim1P_1.fastq.gz",
                read2 = OUTdir + "/trimming/{locust}_trim2P_2.fastq.gz",
                database = KAIJUdir + "/kaiju_db_nr_euk.fmi",
                taxonid = KAIJUdir + "/nodes.dmp",
                taxonnames = KAIJUdir + "/names.dmp",
        output:
                report = OUTdir + "/kaiju/{locust}_kaiju.out",
                classification = OUTdir + "/kaiju/{locust}_kaiju.tsv",        
    shell:
                """
                ml GCC/8.3.0  OpenMPI/3.1.4 Kaiju/1.7.3-Python-3.7.4
                kaiju -z 12 -v -a greedy -f {input.database} -t {input.taxonid} -i {input.read1} -j {input.read2} -o {output.report}
        kaiju2table -v -t {input.taxonid} -n {input.taxonnames} -r phylum -o {output.classification} {output.report}
        """
                
        
########################################
# Parameters in the cluster.json file
########################################

    "kaiju":
    {
        "cpus-per-task" : 6,
        "partition" : "medium",
        "ntasks": 2,
        "mem" : "200G",
        "time": "0-12:00:00"
   },

The ouput produced here allows us to see the percentage of reads that map to unclassified (likely our locust host here) and the percentage of microbial contamination ranked in a phylum level.

Visualize the metatranscriptomics result

Kaiju output can be exported to be view in a interactive and hierarchical multi-layered pie-charts using Krona. The results are generated by a .html page. We followed the Kaiju tutorial on the Github page:

########################################
# Snakefile rule
########################################

rule krona:
        input:
                kaijuout = OUTdir + "/kaiju/{locust}_kaiju.out",
                taxonid = KAIJUdir + "/nodes.dmp",
        taxonnames = KAIJUdir + "/names.dmp",
        output:
                conversion = OUTdir + "/kaiju/{locust}_krona.out",
        webreport = OUTdir + "/kaiju/{locust}_krona.html",
        shell:
                """
                ml GCCcore/8.2.0 KronaTools/2.7.1
                kaiju2krona -t {input.taxonid} -n {input.taxonnames} -i {input.kaijuout} -o {output.conversion}
        ktImportText -o {output.webreport} {output.conversion} 
                """

        
########################################
# Parameters in the cluster.json file
########################################

    "krona":
    {
        "cpus-per-task" : 2,
        "partition" : "short",
        "ntasks": 1,
        "mem" : "500M",
        "time": "0-0:10:00"
   },

Ribosomal RNA

sessionInfo()

FALSE R version 4.3.1 (2023-06-16)
FALSE Platform: x86_64-apple-darwin20 (64-bit)
FALSE Running under: macOS Sonoma 14.1.2
FALSE 
FALSE Matrix products: default
FALSE BLAS:   /Library/Frameworks/R.framework/Versions/4.3-x86_64/Resources/lib/libRblas.0.dylib 
FALSE LAPACK: /Library/Frameworks/R.framework/Versions/4.3-x86_64/Resources/lib/libRlapack.dylib;  LAPACK version 3.11.0
FALSE 
FALSE locale:
FALSE [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
FALSE 
FALSE time zone: America/Chicago
FALSE tzcode source: internal
FALSE 
FALSE attached base packages:
FALSE [1] stats     graphics  grDevices utils     datasets  methods   base     
FALSE 
FALSE other attached packages:
FALSE  [1] ggConvexHull_0.1.0 reshape2_1.4.4     ggthemes_5.0.0     plotly_4.10.4     
FALSE  [5] DT_0.31            lubridate_1.9.3    forcats_1.0.0      stringr_1.5.1     
FALSE  [9] dplyr_1.1.4        purrr_1.0.2        readr_2.1.5        tidyr_1.3.1       
FALSE [13] tibble_3.2.1       ggplot2_3.4.4      tidyverse_2.0.0    rmdformats_1.0.4  
FALSE [17] knitr_1.45         workflowr_1.7.1   
FALSE 
FALSE loaded via a namespace (and not attached):
FALSE  [1] gtable_0.3.4      xfun_0.41         bslib_0.6.1       htmlwidgets_1.6.4
FALSE  [5] processx_3.8.3    callr_3.7.3       tzdb_0.4.0        vctrs_0.6.5      
FALSE  [9] tools_4.3.1       ps_1.7.6          generics_0.1.3    fansi_1.0.6      
FALSE [13] highr_0.10        pkgconfig_2.0.3   data.table_1.15.0 lifecycle_1.0.4  
FALSE [17] farver_2.1.1      compiler_4.3.1    git2r_0.33.0      munsell_0.5.0    
FALSE [21] getPass_0.2-4     httpuv_1.6.14     htmltools_0.5.7   sass_0.4.8       
FALSE [25] yaml_2.3.8        lazyeval_0.2.2    crayon_1.5.2      later_1.3.2      
FALSE [29] pillar_1.9.0      jquerylib_0.1.4   whisker_0.4.1     cachem_1.0.8     
FALSE [33] tidyselect_1.2.0  digest_0.6.34     stringi_1.8.3     bookdown_0.37    
FALSE [37] labeling_0.4.3    rprojroot_2.0.4   fastmap_1.1.1     grid_4.3.1       
FALSE [41] colorspace_2.1-0  cli_3.6.2         magrittr_2.0.3    utf8_1.2.4       
FALSE [45] withr_3.0.0       scales_1.3.0      promises_1.2.1    timechange_0.3.0 
FALSE [49] rmarkdown_2.25    httr_1.4.7        hms_1.1.3         evaluate_0.23    
FALSE [53] viridisLite_0.4.2 rlang_1.1.3       Rcpp_1.0.12       glue_1.7.0       
FALSE [57] formatR_1.14      rstudioapi_0.15.0 jsonlite_1.8.8    R6_2.5.1         
FALSE [61] plyr_1.8.9        fs_1.6.3

Transcriptome sequencing data pre-processing

Maeva Techer

2024-02-19