breast_cancer_recurrence_classifier

Our breast cancer study published on PNAS indicated potential application of SILVER-Seq on breast cancer and identified some promising breast cancer related diagnostic panels, for cancer and recurrence prediction. To develop and further validate the model, we've collected serum samples from PNAS paper and 2 additional sources and processed the samples with SILVER-Seq (as shown below).

Project goals

As stated in the section above, the goals of this project are

• 1. Develop the breast cancer classifier model from PNAS paper using the PNAS breast cancer and normal datasets.
• 2. Develop the breast cancer recurrence classifier model from PNAS paper using the PNAS breast cancer dataset.
• 3. (Optional) Try to apply the above models to the "validation" datasets. (Note: we do not expect a high performance of most models in the "validation" datasets.)

PNAS Samples and Data

Raw Data

Concepts

• poiseid: A unique ID assigned to each donor. There might be several samples from one single donor, usually collected at different timepoints, samples from one donor shared the same poiseid
• fu: Short for “follow-up”. There might be several samples from one single donor, usually collected at different timepoints, samples collected at different timepoints are assigned with different fu number
• sample_id: A unique ID given to data for each sample to connect data with metainfo of the corresponding sample.

Documents

• pnas_counts_96_nodup.txt: Exon level reads counts for all 96 breast cancer samples in PNAS paper. Column 1-96 (S01-S96) in the file corresponds to row 1-96 (S01-S96) in “pnas_patient_info.csv”
• pnas_tpm_96_nodup: Exon level TPM for all 96 breast cancer samples in PNAS paper. Column 1-96 (S01-S96) in the file corresponds to row 1-96 (S01-S96) in “pnas_patient_info.csv”
• pnas_patient_info.csv: Metadata for all 96 breast cancer samples in PNAS paper, row 1-96 corresponds to S01-S96. Column 1-96 in “pnas_counts_96_nodup.txt” and “pnas_tpm_96_nodup” corresponds to row 1-96 in “pnas_patient_info.csv”.
• pnas_normal_tpm.txt: Exon level TPM for all 32 normal samples in PNAS paper.
• pnas_normal_readcounts.txt: Exon level reads counts for all 32 normal samples in PNAS paper.

Validation Samples and Data

Raw Data

Concepts

• Mapping ID (e.g. BCSC_S5_01): A unique ID given to data for each sample to connect data with metainfo of the corresponding sample

Documents

• validation_bc_meta.xlsx: Metainfo for validation breast cancer samples based on their Mapping ID.
• validation_normal_meta.xlsx: Metainfo for SDBB validation normal samples based on their Mapping ID
• validation_exon_readcounts: Exon level read counts for validation breast cancer and normal samples based on their Mapping ID.
• validation_exon_tpm: Exon level TPM for validation breast cancer and normal samples based on their Mapping ID.