seq2scfv-fixes

2025-11-29 Update

The fixes described in this repository are no longer sufficient. A fully updated and functional version of the Dockerfile will is now maintained here: https://github.com/MikolajKocikowski/seq2scfv-unofficial-updated. This repository is retained for archival purposes.

About

Seq2scfv is a fantastic tool for scFv analysis, but I had trouble getting it to run on my machines. This repository documents fixes and workarounds that made it functional for me. It's not affiliated with the original authors. I submitted a related issue and pull requests, but given that the official repo is tied to a published paper, I understand that updates may be limited. Hence, I will maintain any further updates or tweaks here in case they're useful to others. I'm not a Docker expert - feedback always welcome!

The Issue

I encountered compatibility and permission issues when building the Dockerfile with R 3.6.3, which is the default R version in Ubuntu 20.04 (the base image used in this Dockerfile). Specifically:

1. ggplot2 installation fails due to the unavailability of gtable for R 3.6.3 on CRAN.
2. A permissions issue prevents install2.r from executing successfully.

The Solutions

There are two versions of the solution; both modify a chunk of the Dockerfile code (make it look like below).

Simple Solution: This solution installs gtable version 0.3.0 from the CRAN archive, which is compatible with R 3.6.3, and sets the correct permissions for install2.r. This approach ensures the build completes successfully without many changes.

# Ensure install2.r has executable permissions
RUN chmod +x /usr/bin/install2.r

# Install gtable version 0.3.0 from the CRAN archive
RUN Rscript -e "install.packages('https://cran.r-project.org/src/contrib/Archive/gtable/gtable_0.3.0.tar.gz', repos = NULL, type = 'source')"

# Install other necessary R packages, including ggplot2 and gridExtra
RUN /usr/bin/install2.r --libloc /home/R --error --repos "https://stat.ethz.ch/CRAN/" \
    data.table \
    reshape2 \ 
    dplyr \
    tidyr \
    ggplot2 \
    gridExtra \
    stringr

Alternative Solution: If you prefer the newest acceptable gtable version (0.3.5), this solution seems to work for me, but requires separating the installation of the libraries into a few steps.

# Set executable permission for install2.r
RUN chmod +x /usr/bin/install2.r

# Install core libraries first to cache this layer
RUN /usr/bin/install2.r --libloc /home/R --error --repos "https://stat.ethz.ch/CRAN/" \
    data.table \
    reshape2 \
    dplyr \
    tidyr \
    stringr

# Install gtable version 0.3.5 from the archive
RUN Rscript -e "install.packages('https://cran.r-project.org/src/contrib/Archive/gtable/gtable_0.3.5.tar.gz', repos = NULL, type = 'source')"

# Install ggplot2 and gridExtra separately to prevent conflicts
RUN /usr/bin/install2.r --libloc /home/R --error --repos "https://stat.ethz.ch/CRAN/" \
    ggplot2 \
    gridExtra

Summary

I hope this helps anyone trying to replicate the setup or adapting the environment, especially as CRAN packages evolve over time. Perhaps switching the R version would be easier, but it would impact reproducibility and I learned not to mess with R in the past!

Internally, seq2scfv depends on IgBlast, which can be tricky to configure. The documentation is limited, and errors are common - especially when working with non-default species that require custom-prepared reference files. I've seen issues raised about it without clear solutions.

After a fair bit of trial, error, and reverse engineering, I'm preparing a tutorial on such IgBlast setup - stay tuned.

I also plan to release tools for analyzing and visualizing seq2scfv output - links coming.

If you're interested in these, please get in touch, so we can discuss the needs, features, or collaborate on it!