#connect to server

ssh root@10.19.139.110
stu-25617

#Convert metadata file type
qiime metadata tabulate \
  --m-input-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization infant_metadata.qzv

#navigate to dataset
cd /
cd mnt/datasets/project_2/infant

#Use the manifest to import as a .qza file
qiime tools import \
  --type "SampleData[SequencesWithQuality]" \
  --input-format SingleEndFastqManifestPhred33V2 \
  --input-path ./infant_manifest.txt \
  --output-path /data/infant_demux_seqs.qza

cd /
cd data

qiime demux summarize \
  --i-data ./infant_demux_seqs.qza \
  --o-visualization ./infant_demux_seqs.qzv

#Denoise the sequences
qiime dada2 denoise-single \
  --i-demultiplexed-seqs ./infant_demux_seqs.qza \
  --p-trunc-len 150 \
  --o-table ./infant_dada2_table.qza \
  --o-representative-sequences ./dada2_rep_set.qza \
  --o-denoising-stats ./infant_dada2_stats.qza

#Convert denoising stats to visualizable file
qiime metadata tabulate \
  --m-input-file ./infant_dada2_stats.qza  \
  --o-visualization ./infant_dada2_stats.qzv

#Generate feature table summary
qiime feature-table summarize \
  --i-table ./infant_dada2_table.qza \
  --m-sample-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization ./infant_dada2_table.qzv

#Download SILVA database
wget \
  -O "sepp-refs-silva-128.qza" \
  "https://data.qiime2.org/2021.8/common/sepp-refs-silva-128.qza"

#Generate the phylogenetic tree 
qiime fragment-insertion sepp \
  --i-representative-sequences ./dada2_rep_set.qza \
  --i-reference-database /data/sepp-refs-silva-128.qza \
  --o-tree ./infant_tree.qza \
  --o-placements ./infant_tree_placements.qza \
  --p-threads 8


#Filter out adult samples
qiime feature-table filter-samples \
  --i-table infant_dada2_table.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --p-where "[life_stage]='Infant'" \
  --o-filtered-table infant-filtered-table.qza


#Feature table for filtered file
qiime feature-table summarize \
  --i-table ./infant-filtered-table.qza \
  --m-sample-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization ./infant-filtered-table.qzv


#Generate alpha rarefaction curve
qiime diversity alpha-rarefaction \
  --i-table ./infant-filtered-table.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization ./infant-filtered_alpha_rarefaction_curves.qzv \
  --p-min-depth 10 \
  --p-max-depth 30000


#Generate alpha rarefaction curve - but KEEPING ALL INFANT YES PROB SAMPLES (MAX SAMPLING DEPTH 492)
qiime diversity alpha-rarefaction \
  --i-table ./infant-filtered-table.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization ./infant-filtered_alpha_rarefaction_curves.qzv \
  --p-min-depth 10 \
  --p-max-depth 492

#Export data
scp root@10.19.139.110:/data/infant-filtered_alpha_rarefaction_curves.qzv .

# Taxonomic analysis
qiime feature-classifier classify-sklearn \
  --i-classifier /mnt/datasets/classifiers/silva-138-99-515-806-nb-classifier.qza \
  --i-reads dada2_rep_set.qza \
  --o-classification taxonomy.qza

qiime metadata tabulate \
  --m-input-file taxonomy.qza \
  --o-visualization taxonomy.qzv

#Taxabar plots
qiime taxa barplot \
  --i-table infant-filtered-table.qza \
  --i-taxonomy taxonomy.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization infant-taxa-bar-plots.qzv

#Perform diversity analyses 
qiime phylogeny align-to-tree-mafft-fasttree \
  --i-sequences dada2_rep_set.qza \
  --o-alignment aligned-rep-seqs.qza \
  --o-masked-alignment masked-aligned-rep-seqs.qza \
  --o-tree unrooted-tree.qza \
  --o-rooted-tree rooted-tree.qza

qiime diversity core-metrics-phylogenetic \
 --i-table ./infant-filtered-table.qza \
 --i-phylogeny ./infant_tree.qza \
 --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
 --p-sampling-depth 5000 \
 --output-dir ./core-metrics-results

#Core metrics using lower sampling depth to keep all infant yes probiotic samples
qiime diversity core-metrics-phylogenetic \
 --i-table ./infant-filtered-table.qza \
 --i-phylogeny ./infant_tree.qza \
 --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
 --p-sampling-depth 70 \
 --output-dir ./core-metrics-results_aim1

#Visualizing alpha diversity files
qiime diversity alpha-group-significance \
  --i-alpha-diversity ./core-metrics-results_aim1/faith_pd_vector.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization ./core-metrics-results_aim1/faiths_pd_statistics.qzv

#Visualizing beta diversity files - PERMANOVA
qiime diversity beta-group-significance \
  --i-distance-matrix core-metrics-results_aim1/unweighted_unifrac_distance_matrix.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --m-metadata-column probiotic_inf \
  --o-visualization core-metrics-results_aim1/unweighted-unifrac-prob_inf-significance.qzv

qiime diversity beta-group-significance \
  --i-distance-matrix core-metrics-results_aim1/weighted_unifrac_distance_matrix.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --m-metadata-column probiotic_inf \
  --o-visualization core-metrics-results_aim1/weighted-unifrac-prob_inf-significance.qzv


#Filter out adult samples
qiime feature-table filter-samples \
  --i-table infant_dada2_table.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --p-where "[life_stage]='Infant'" \
  --o-filtered-table infant-filtered-table.qza

#Filter out not collected samples (for infant probiotic use)
qiime feature-table filter-samples \
  --i-table infant-filtered-table.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --p-where "[probiotic_inf]='yes' OR [probiotic_inf]='no' " \
  --o-filtered-table infant-filtered2-table.qza


#Feature table for filtered file
qiime feature-table summarize \
  --i-table ./infant-filtered2-table.qza \
  --m-sample-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization ./infant-filtered2-table.qzv

#Export to local computer
scp root@10.19.139.110:/data/infant-filtered2-table.qzv .

#Generate alpha rarefaction curve - KEEPING ALL INFANT YES PROB SAMPLES (MAX SAMPLING DEPTH 492)
qiime diversity alpha-rarefaction \
  --i-table ./infant-filtered2-table.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization ./infant-filtered2_alpha_rarefaction_curves.qzv \
  --p-min-depth 10 \
  --p-max-depth 492


#Taxabar plots
qiime taxa barplot \
  --i-table infant-filtered2-table.qza \
  --i-taxonomy taxonomy.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization infant-filtered2-taxa-bar-plots.qzv

#Diversity metrics
qiime diversity core-metrics-phylogenetic \
 --i-table ./infant-filtered2-table.qza \
 --i-phylogeny ./infant_tree.qza \
 --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
 --p-sampling-depth 70 \
 --output-dir ./core-metrics-results-aim1_filtered2

#Visualizing alpha diversity files
qiime diversity alpha-group-significance \
  --i-alpha-diversity ./core-metrics-results-aim1_filtered2/faith_pd_vector.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization ./core-metrics-results-aim1_filtered2/faiths_pd_statistics.qzv

scp root@10.19.139.110:/data/core-metrics-results-aim1_filtered2/faiths_pd_statistics.qzv .

# no error after filtering out not collected
qiime diversity alpha-group-significance \
 --i-alpha-diversity ./core-metrics-results-aim1_filtered2/evenness_vector.qza \
 --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
 --o-visualization ./core-metrics-results-aim1_filtered2/evenness_statistics.qzv

#Visualizing beta diversity files - PERMANOVA
qiime diversity beta-group-significance \
  --i-distance-matrix core-metrics-results-aim1_filtered2/unweighted_unifrac_distance_matrix.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --m-metadata-column probiotic_inf \
  --o-visualization core-metrics-results-aim1_filtered2/unweighted-unifrac-prob_inf-significance.qzv

qiime diversity beta-group-significance \
  --i-distance-matrix core-metrics-results-aim1_filtered2/weighted_unifrac_distance_matrix.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --m-metadata-column probiotic_inf \
  --o-visualization core-metrics-results-aim1_filtered2/weighted-unifrac-prob_inf-significance.qzv




#Filter out not collected samples (for mother use of probiotic)
qiime feature-table filter-samples \
  --i-table infant-filtered-table.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --p-where "[probiotic_mom]='yes' OR [probiotic_mom]='no' " \
  --o-filtered-table infant-filtered3-table.qza

#Feature table for filtered file
qiime feature-table summarize \
  --i-table ./infant-filtered3-table.qza \
  --m-sample-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization ./infant-filtered3-table.qzv

qiime diversity alpha-rarefaction \
  --i-table ./infant-filtered3-table.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization ./infant-filtered3_alpha_rarefaction_curves.qzv \
  --p-min-depth 10 \
  --p-max-depth 1518

#Diversity metrics
qiime diversity core-metrics-phylogenetic \
 --i-table ./infant-filtered3-table.qza \
 --i-phylogeny ./infant_tree.qza \
 --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
 --p-sampling-depth 195 \
 --output-dir ./core-metrics-results-aim1_filtered3

#Visualizing alpha diversity files
qiime diversity alpha-group-significance \
  --i-alpha-diversity ./core-metrics-results-aim1_filtered3/faith_pd_vector.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization ./core-metrics-results-aim1_filtered3/faiths_pd_statistics.qzv

scp root@10.19.139.110:/data/core-metrics-results-aim1_filtered3/faiths_pd_statistics.qzv .

# no error after filtering out not collected
qiime diversity alpha-group-significance \
 --i-alpha-diversity ./core-metrics-results-aim1_filtered3/evenness_vector.qza \
 --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
 --o-visualization ./core-metrics-results-aim1_filtered3/evenness_statistics.qzv

#Visualizing beta diversity files - PERMANOVA
qiime diversity beta-group-significance \
  --i-distance-matrix core-metrics-results-aim1_filtered3/unweighted_unifrac_distance_matrix.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --m-metadata-column probiotic_mom \
  --o-visualization core-metrics-results-aim1_filtered3/unweighted-unifrac-prob_inf-significance.qzv

qiime diversity beta-group-significance \
  --i-distance-matrix core-metrics-results-aim1_filtered3/weighted_unifrac_distance_matrix.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --m-metadata-column probiotic_mom \
  --o-visualization core-metrics-results-aim1_filtered3/weighted-unifrac-prob_inf-significance.qzv


#Filtering for aim 2 - similar probiotic treatment intake patterns

#Filter out adult samples
qiime feature-table filter-samples \
  --i-table infant_dada2_table.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --p-where "[life_stage]='Infant'" \
  --o-filtered-table infant-filtered-table.qza


# Filtering for subjects 1, 14 and 27 - only had probiotics at 2 months

qiime feature-table filter-samples \
  --i-table infant-filtered-table.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --p-where "[anonymized_name]='70001' OR [anonymized_name]='70014' OR [anonymized_name]='70027'" \
  --o-filtered-table infant-1and14and27-table.qza

qiime feature-table summarize \
  --i-table ./infant-1and14and27-table.qza \
  --m-sample-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization ./infant-1and14and27-table.qzv

qiime diversity alpha-rarefaction \
  --i-table ./infant-1and14and27-table.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization ./infant-1and14and27_alpha_rarefaction_curves.qzv \
  --p-min-depth 10 \
  --p-max-depth 15513

#Taxabar plots
qiime taxa barplot \
  --i-table infant-1and14and27-table.qza \
  --i-taxonomy taxonomy.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization infant-1and14and27-taxa-bar-plots.qzv

#Core diversity metrics
qiime diversity core-metrics-phylogenetic \
 --i-table ./infant-1and14and27-table.qza \
 --i-phylogeny ./infant_tree.qza \
 --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
 --p-sampling-depth 1900 \
 --output-dir ./core-metrics-results-infant-1and14and27

#Visualizing alpha diversity files
qiime diversity alpha-group-significance \
  --i-alpha-diversity ./core-metrics-results-infant-1and14and27/faith_pd_vector.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --o-visualization ./core-metrics-results-infant-1and14and27/faiths_pd_statistics.qzv

scp root@10.19.139.110:/data/core-metrics-results-infant-1and14and27/faiths_pd_statistics.qzv .

qiime diversity alpha-group-significance \
 --i-alpha-diversity ./core-metrics-results-infant-1and14and27/evenness_vector.qza \
 --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
 --o-visualization ./core-metrics-results-infant-1and14and27/evenness_statistics.qzv

#Visualizing beta diversity files - PERMANOVA
qiime diversity beta-group-significance \
  --i-distance-matrix core-metrics-results-infant-1and14and27/unweighted_unifrac_distance_matrix.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --m-metadata-column age_category \
  --o-visualization core-metrics-results-infant-1and14and27/unweighted-unifrac-prob_inf-significance.qzv

qiime diversity beta-group-significance \
  --i-distance-matrix core-metrics-results-infant-1and14and27/weighted_unifrac_distance_matrix.qza \
  --m-metadata-file /mnt/datasets/project_2/infant/infant_metadata.txt \
  --m-metadata-column age_category \
  --o-visualization core-metrics-results-infant-1and14and27/weighted-unifrac-prob_inf-significance.qzv